Local Analysis of Air Quality Changes in the Community of Madrid before and during the COVID-19 Induced Lockdown

Round 1

Reviewer 1 Report

Review report:

I think that the authors did an important effort to improve the paper from the first version. The actual paper still has some point to address. They are listed below.

General comments

abstract: the abstract is too long and a bit difficult to read in this new form!! There is too many information, it must be summarized. For example, L42 to 79 a list of papers, that are then summarized in L80 to 87….

In my opinion, these three 3 paragraphs should be group in one.

Then L98-130: I think it could a good idea to highlight what distinguish the paper from the existing literature, but in my opinion, such a level of details is not necessary (this is more to convince the editor or a reviewer than appropriated to an introduction of a scientific paper …) I think that the same can be said in less sentences.

L.135 – L.152 present some results while we are still in the introduction section. That has no sense… The results should only be presented in the appropriate section or partly in the abstract.

Section 3.5: I think that the authors should comment the significant increase of O3 in Madrid (see Fig 6a), to be considered in the context of the air quality: A strong decrease in NOx emissions will lead to a significant increase in O3.

L738-742: what the authors are talking about is secondary aerosols. In particular secondary organic aerosols are important in the context of an urban area. They can arise from VOCs emitted by many sources. Please add some words about secondary aerosols in your discussion.

L765 – 770: I do not fully agree with the authors saying that ground-based measurements are more representative than satellite approaches. Both are complementary! Ground based measurement only measure the air quality at 1 point: This does not allow to consider the spatial variability of pollutants, while the satellite measurement integrate all the CM.

Please rephrase these last sentences.

 

minor comments

L9-10. replace « pollution particles » by « pollutants».

96 and L333-334: « the strictest quarantine regime in Europe ». What are the criteria to compare all the regime in Europe? I think “one of the strictest” would be more appropriated …

Figure 1-b: “type os station” -> “type of”

Table 1: legend must be more explicit.

569 to 571: I would like to find these sentences in the legend of fig 5 and 6 rather than in the main text.

L610: Figure 5 -> figure 6

682 – 685: This is more appropriated to introduce the method of cluster analysis than results.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

In the revised manuscript, authors took into consideration some basic comments from my side. They included 1-2 new figures and extend the discussions in some parts of the manuscript. However, I believe that the new figures are rather worst than the previous ones, and at least, less readable. So, I cannot understand why authors replaced the previous ones with these new. On the other hand, I recommended for such an important issue that has attracted the global interest with numerous published works till now about the COVID-19 lockdown impact on air quality and aerosols, authors to increase the literature and comparison with previous works at least in Spain, and for the similar environments in the Mediterannean (i.e. Greece, Italy, etc) and/or Europe. On the contrary, in the revised manuscript, the references have been reduced about 10, compared to the original submission...   

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The paper analyses the effect of lockdown and Low Emission Zone in the Spanish Community of Madrid, by considering data from 43 air quality monitoring stations and meteorological parameters. The topic is very interesting and the paper is well written and organized. However, I have few comments, questions and suggestions to improve the overall quality of the manuscript.

Introduction:

Page 1 line 40-41: in order to complement/support the sentence, I suggest the authors to add specific data about traffic reductions during Covid-19 lockdown in the considered study area.

The authors should expand this section by including papers where similar analyses were developed; for this reason, I suggest to consider further journals, like, but not limited to, those from the current publisher (e.g. Sustainability).

Methodology:

Page 4 line 179-180: the authors declared that there are seven types of monitoring stations, however, in the following, just six types are listed and in Figure 1 there are just five types. Why?

Page 5: typos on legend of Figure 1b

Page 7 line 30: How did the authors define the grid dimensions?

I suggest the authors to organize the paragraphs to follow the same order in the Results section, thereby helping the reader to understand the developed analysis.

Results:

Page 13 line 525-533: adding data quantifying the reduction/increase of traffic in the area would be useful to support the sentences.

Page 15 line 568: the authors should explain/recall the meaning of percentages and values in brackets in Figure 5.

Page 15 line 608-310: I suggest the authors to discuss about the following questions to highlight their contribution on the topic: why an econometric model was needed to study the effect of LEZ and LD on different sites? Could the same considerations be directly obtained by analyzing air quality data?

Page 15 line 591: the authors should clarify why NO2 concentrations in nearby surrounding areas of Madrid decreases even if traffic congestion increased (page 13 line 525).

Conclusions:

In page 15 lines 589-592 the authors stated that “Thus, for instance, panel g) of Figure 5 shows that the creation of Madrid's LEZ resulted in large decreases of NO2 concentration in nearby surrounding areas of Madrid. Such decreases range from -0.2% to -20.2%, and turn positive in only three cases.”. However, in page 22 lines 793-795, they wrote “The results also showed that the effect of Madrid's low emission zone (the so-called Madrid Central) upon NO2 within the city was significant, between -10% and 20%, and that surrounding areas also benefited from the initiative.”. In the first case, the reduction was referred to surrounding zones, whereas in the second case, the reduction was referred to the area within the city. I suggest the authors to rephrase the reported sentences in the Conclusions section in order to avoid misunderstandings.

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 4 Report

This study documents the COVID lockdown impacts on multiple pollutants over the Spanish Community of Madrid. The work relies largely on surface observations at 40+ sites in the study region. It is an addition to many existing studies on the same topic, I find the paper lacks novelty and its main contributions are overstated. More importantly, the methods they applied have obvious limitations which reduce the robustness of their results. Specifically

1) The pre-lockdown time period (2018/2019) they studied is too short for developing climatology. Because the ground based sites were established way before 2018, why not develop the analysis based on a longer data record?

2) there lacks introduction to method of pollutants (and their associated uncertainty, this is especially important to quantitative NO2 analysis as some methods may lead to positive biases) and weather data (where are the sites?). Are there any VOC data? At line 179, they stated 7 kinds of stations but only specified 6.

3) Ozone analysis is too coarse. The analysis is largely relied on daily averages, not linked to policy-relevant metrics, exceedances and no explanations based on chemical regimes were provided. 

4) The authors argue that their methods are better than satellite observations, ignoring that satellite data have much better spatial coverage and offer additional information on chemical regimes, such as NO2/HCHO ratio, which help understand ozone chemistry better. I found that the citations of satellite-based analysis of this topic are lacking. Their results should be compared and discussed together with satellite-based studies.

Finally, the paper has structural and presentation shortcomings. For example, section 2.5 should be moved forward. And also, Figure 2 is very difficult to understand. The authors should at least consider showing difference plots for each of these species. The authors should conduct a more thorough literature review to strengthen their results discussions and avoid overstating their contributions.

These issues of the paper will most likely not be addressed after a major revision, I therefore suggest the paper to be rejected at this time. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 5 Report

The manuscript “Local Analysis of Air Quality Changes in the Community of Madrid before and during the COVID-19 Induced Lockdown” by Betancourt-Odio et al. examines the effect of the COVID-19 induced lockdown and of other variables (including anti-pollution traffic policies) in the air quality of the city of Madrid and the sorrounding areas. The subject is really interesting in this period. The paper adopted a multivariate approach, for controlling various factors that could influence air quality, at the same time, instead of simply compare pollution levels during the pandemic period with reference previous periods. The manuscript is well written; the introductin fully describes the state of the art, methods are comprehensively described and conclusions are supported by the data.

In my opinion, it should be published after minor revisions.

Minor comments:

• Lines 10-11. Sobstitute “pollution particles” with “pollutants”, since not all pollutants are present in form of atmospheric particles (CO, NO, NO₂, O₃).
• Line 64. Delete “a”.
• Lines 89. Sobstitute “pollution particles” with “pollutants”.
• Lines 140-152. It seems to me that these are the results of the paper. Delete it from the introduction.
• Lines 153-156. I have some doubts in inserting here the description of the sections.
• Figure 1 (lines 190-193), Figure 5 (Lines 623-631), Figure 6 (lines 653-662), Figure 7 (line 707). I think it should be better to enumerate the monitoring stations, so that the same station is associated to the same number throughout the article.
• Lines 195 (caption Figure 1), lines 633-635 (caption Figure 5), lines 665-667 (caption Figure 6). I suggest to use the same nomenclature in the caption for referring to the sorrounding area of Madrid: “CM” or “Madrid sorrounding area”.
• Line 206. Delete “_”.
• Line 217. Insert “episodes” instead of “episode”.
• Line 311. Insert “grid” instead of “Grid”.
• Figure 3 (line 438). I suggest to insert horizontal lines for highlighting the average of the various periods (before LEZ, after LEZ and during LD), in different colours.
• Line 598. “Figure 5” instead of “Figure 4”.
• Line 610. I suppose it was Figure 6.
• Line 621. Insert “air pollutants” instead of “particles”.
• Line 691. As I can see in my version, cluster 2 is green, not pink.
• Lines 701-705. I suggest to delete this part, since it is more suited for the “discussion” or the “conclusion” section. I observed that this concept is already mentioned in the conclusions.
• Line 719. Insert the references “Chauhan and Singh” and “Sicard et al.” as numbers in square brackets.
• Line 736. Subscript “10” pf “PM₁₀”.
• Line 740. Insert a comme before “but”.
• Line 744. Insert “and” instead of “y”.
• Line 783. Insert “pollutants” instead of “particles”.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Authors provided some edits mostly in abstract and introduction sections that somewhat improved the manuscript without changing its importance or significance. However, tha narrative flows well, the figures have been improved and the paper can be accepted, however, without being considered as a significant contribution to science. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This paper is investigating the influence of the Covid-19 lockdown on the several criteria pollutants in Madrid. As expected, when compared with the pre-lockdown period, it is found that NOx and PM exhibit reduction in concentrations during the lockdown period. There are a few major issues with this data analysis: 1) the lockdown values (March-June 2020) were compared to pre-lockdown values (Jan-March 2020, a period when traffic restriction was put in place). It should make more sense to compare March-June 2020 to March-June 2018 and 2018. To investigate the influence of traffic restriction on air quality, the authors could compare Jan-March 2020 to Jan-March 2018 and 2019. Trafic restriction and its influence on air quality has been discussed in the manuscript, but there was no attempt to quantify that. 2) NOx, NO2 and NO are presented as different pollutants. As NOx is NO + NO2 I suggest restructuring figures and text and present NO and NO2 or NOx. 3) it should be investigated whether this dataset is long enough for the meteorological data to be taken into account – i.e. to perform meteorological normalisation (as in for example: https://doi.org/10.1016/j.atmosenv.2020.118048) 2)it is not clear what colour scheme in Figures 3 and 4 is showing. After addressing these and comments below, I think that the manuscript can be considered for publication in the Atmosphere

 

Line 17 and the whole paper: Whey are NOx, NO2 and NO presented and discussed separately? NOx is NO + NO2, so you should either present NOx or NO and NO2. Related to this in line 14 should be six pollutants, not seven.

Line 29: should be pollutants

Line 32: pollutants’ emissions

Line 35: instead of “matter interaction” I suggest interaction of pollutants

Line 37: what do you mean by consumption?

Line 37: you mean human health?

Line 53: NO2 is the major component of NOx, so it is expected that they will have the same trend

Line 58-60: reference needed for this sentence

Line 78: what do you mean by calendar effects? Weekday/weekend?

Line 79: debate

Line 84:  identified instead of “detected precisely”

Line 93: Please fix this sentence – design(develop) effective policies is repeated twice.

Line 96-100: please remove as the paper follows the standard journal manuscript format.

Line 100: Can be called just Methods or Methodology (as you don’t have any Materials). Also, as you are reporting results related to travel restriction period, somewhere in the Methodology you should put in a couple of sentences regarding traffic restriction – when was it introduced, why and what did it involve.

Line 105:I think that this sentence is not relevant for the paper and can be removed

Line 107: no need to convert into square miles  - SI units are used in scientific publishing/communication, not imperial.

Line 129: Figure 1: It would be good if stations were coloured based on what kind it is: urbat, rural, etc.

Line 146: R is a programming language and software that uses various packages to manipulate/visualise data, so it is not correct tot say R statistical package – instead should be: R software using the following statistical packages:…

Line 146: what do you mean by marginal impact?

Line 151: How was influence of meteorological factors in 2018, 2019 and 2020 taken into account? How do you know that any differences that you’ve observed aren’t due to different met. Conditions? You should describe in the manuscript how you’ve done meteorological normalisation.

Line 186: using instead of means

Line 194: it says here that pollution maps show average pollutant concentrations in Jan-March 2018-2019 and line 272 says that pollution maps show average pollutant concentrations in the same period as the lockdown , but in pre-covid years. Which one is correct?

Line 267: again, how were met conditions taken into account?

Line 280: Based on the regression model’s results ……..and PM2.5 (Figure 3 and 4)

Line 284: What are primary pollutants? Please identify.

Line 287: “difference” instead of “gap”?

Line 288: which secondary pollutants? You mean O3?

Line 310: before lockdown or in March-June 2018-2019?

Line 314 – this is where discussion of Fig 3 starts. Figure 3 needs to be introduced here and explained what is being compared and why? The way it is currently makes it hard to follow as you first compare lockdown comparing to the same months in 2018 and 2019 and then in Fig 3 you compare pre-covid period (Jan-March), when Madrid had traffic restrictions to the covid lockdown

Line 316 (and throughout the manuscript) – please correct ug to µg

Line320-323: I believe that you want to say here that you were expecting more substantial decrease in CO during the lockdown, but this was offset by more intense use of house heating appliances during lockdown which started in winter. Please correct your sentence as at the moment it reads as the slight decrease in CO was due more intense use of heating appliances. Also, this hypothesis might be true for March, but not for May or June, so you can easily test your hypothesis by comparing daily CO values in March 2018-19 vs 2020 and May,June 2018-19 vs 2020.

Line 326 – pre-lockdown levels (e.g. Feb 2020) or March-June of pre-covid years (2019-2020)?

Line 339: remove “pollutants”. The way it is written sounds like there are several NO2 pollutants

Line354: when mentioning these stations, please describe them as either rural or urban. Are those increases/decreases significant?

Line 366: is a drop of 3.2 ug/m3 significant?

Line 369: why was the greatest reduction in PM2.5 in rural areas? I would expect PM2.5 to follow the same trend as PM10

Line 404: What do you mean by marginal level/ marginal impact? What does colour legend show? I do not understand those values. They have not been discussed anywhere in the manuscript so it is hard to fully understand this figure (and Fig 4) Also, in the figures TR is on the left side and in the comments it says that it is right figures – please correct that (same comment for Fig 4)

Line 432: It has not been described anywhere how meteorology is controlled in this study

Line 452: As you are trying to identify the influence of the covid-19 lockdown on criteria pollutants’ levels, have you considered excluding intense Saharan dust episodes from your dataset as these events can significantly impact your average values of PM and possibly ozone.

Line 455: have you considered looking at the impact of traffic restrictions on criteria pollutant levels (i.e. comparing Jan-March 2020 with the same period in 2018 and 2019)? Or has that already been done? If yes, you should include it in the discussion

Line 463: Results of the cluster analysis should be in the Results section. Also, there should be more discussion on the meaning of these results. Do they make sense considering spatial arrangement of Madrid’s traffic/industry activities?

Line 465 (and further down on the page): you need to put Figure number when reporting clusters

Line 480: We have not intended instead of We do not pretend.

Line 482: this is contradictory statement as you mentioned a couple of times in the manuscript that meteorology was taken into account (Line 432 and line 152)

Line 487: mention what statistical model

Line 490: here you mention that there was special emphasis on meteorology which is again in contrast with the line 482. How? Emphasis on met is not visible anywhere in the manuscript. Also it is not visible how the effects of the calendar were taken into account/addressed (I am assuming that effect of the calendar is weekdays/weekends)

Line 493 – see my comment for Line 320

Line 501: starting with ..,most likely – I do not understand this part of the sentence – please clarify.

Line 503 – 509: this is a repeat from the discussion – please change. Conclusion should not be repeating things discussed in the Discussion, but provide short, overall summary.

Line 528: are playing instead of may play

Reviewer 2 Report

Review report:

This manuscript by Betancourt-Odio et al. presents a study about the effect of lockdown induced by COVID 19 pandemic on air quality in the community of Madrid.

The authors used a complex dataset from 44 monitoring stations in Madrid area that includes 2 years of measurement before the lockdown and covered 7 different pollutants. The effect of lockdown on the level of pollutants has been characterised using linear regression and clustering methods have been used to characterise the effect of lockdown on air quality, and to compare with the effect traffic restrictions.

The aim of the is clearly introduced and presented, and the paper well-structured and easy to read. However, the discussion can be improved, and some precisions are necessary. In particular, it is not clear through the paper how important is the variability of pollutant concentrations due to meteorological factors. This is important to understand the part of the observed drop of pollutant levels that can be addressed to the natural variations (i.e., meteorology) and to reduced activity/traffic.

I do not have a lot of remarks, but I listed my comments below to try to help the authors to improve their manuscript.

 

General comments

L.131-134 : Measurement stations are monitored either by the municipal air quality network of Madrid city or either by regional authority control What is impact on data ? Is the quality of data similar on each network? Does any bias exist between 1 type of dataset and the other? Are the same instruments used by both air quality monitoring networks?

A full list of instruments of each monitoring station is not required, and I do not doubt about the quality of data used, but more details are needed here to understand what s behind these 2 measurement networks (roughly type of instruments, quality check of data, etc).

L.431 - 432: “The role of meteorology is also evident, and it was controlled in this study.” The role of meteorology is not discussed at all in the paper. In the results section, the authors presented the differences in the meteorological parameters over the years and areas in the CM (L.245 – L.269). After a presentation of the results, one expects that they are discussed. The point is not to provide a quantitative evaluation, but to discuss how meteorological conditions affect the pollutants through the years? If this is evident, as stated by the authors L.431-432, please explain the role of meteorology on pollutant levels. To my mind, the authors must discuss first the effect of meteorological conditions on pollutant levels, then consider the meteorological conditions when the effect of lockdown and traffic restrictions on air quality are discussed. As an example, 2019 was drier than 2018 and 2020, reducing the average air pollutants levels in 2018 and 2020 due to more rain-washing. Thus, at same emissions levels, one can expect higher level of pollutant in 2019…

Section 2.3: The dates of the different lockdown levels are clearly presented. But first, it is necessary to precise the date when traffic restriction (Madrid Central low emissions zone) started. Then, I was wondering whether the authors noticed different pollutant reductions associate to the different lockdown levels (I guess a more important drop of pollutant levels during the harder lockdown?)? This could be helpful for stakeholders to understand the level of restrictions needed to really improve the air quality in CM.

Figure 3 and 4: These 2 figures are interesting, but it is absolutely not clear what’s behind them. I understand how the authors obtained them, but I am questioning the period retained to compute these graphs. For example, is the reference period (the period from which the concentration reduction is obtained) for the TR and lockdown graphs the same? What are the periods used for the means? Are all differences significant (for pollutants and stations)? Please add in the text all the necessary information.

Please reduce number of digits on these graphs, they are not really appropriated.

460 – 462: I would just comment the conclusions of the authors. It Is evident that reduction of economic activity such as during lockdown drastically reduced emissions. However, it should be kept in mind that such a reduction in economic and social activities is not conceivable, because this would lead to a catastrophic situation considering unemployment rate, mental health… I think the drop of pollutants observed during lockdown could be seen as an upper limit of politic of restriction of emissions by simply reducing economic and social activities. It shows that to achieve a good air quality in cities, we need to combine reduction in emissions to other solutions such as cleaner technologies, greener cities, etc.

503 – 502: These lines are a copy paste of L. 445 – 552. Please make new sentences/rephrase.

Comments for author File: Comments.pdf

Reviewer 3 Report

This paper is one of the numerous works that analyze the air quality improvement during the COVID-19 lockdown period. The application area is the Madrid metropolitan and surrounding regions. Authors use classic air pollutants (i.e. NO, NO2, NOx, CO, O3, PM2.5, PM10) from many stations, while they also classify the stations regarding the pollution reductions in 4 clusters. Overall, the paper should be re-arranged mostly in its presentation (too many small paragraphs that are rather redundant), discussions (more analysis is needed for the reductions of air pollutants with inclusion of new graphs) and comparisons with previous studies (literature and findings from other works related to reductions in aerosols and pollutants during the lockdown period is rather poor).  

Abstract should include some basic results regarding the reduction of some pollutants, so the reader to follow them from the beginning.

Literature may be increased for studies dealing with reduction in air pollutants during the COVID-19 period. Also, the lower reductions in PM2.5 levels compared to NOx is a fact that needs further justification with references from studies in Asia, Europe, etc… Same issue for the NO titration effect for the higher O3 levels. For example, you may see studies in the Mediterranean, Europe i.e. Evangeliou et al. (2020 and refs. therein), India (e.g. Dumka et al., 2020, Atm.Poll.Res) and abroad that combined observations from several sites.

Introduction section is composed of several small paragraphs that somewhat provides a confusion. I would suggest to merge some of them, those with similar meaning, thus reducing the number to 4-5 paragraphs, as usually followed in articles. The importance of this work synergistically or against previous ones in Spain dealing with air quality changes under certain periods (like lockdown) should be also highlighted.  

References may be also used for the clustering analysis. Did you use K-means on it? Some more details are needed for this methodology.

The results section should be re-arranged and be divided in sub-sections regarding the changes in meteorology, the changes-reductions in air pollutants. Also, it would be nice authors provide the respective mapping for CO in Fig. 2.

The numerous small paragraphs are rather confusing. Authors should merge the paragraphs into larger ones with the same/similar meaning or discussion of a specific pollutant. Apart from the absolute values in the pollutant reductions, authors should provide the % changes, as well as the statistical significance of these changes, which are the most important. The reduction in air pollutant concentrations in absolute values is of minor importance. This should be followed throughout the results section. The current findings of pollutant reductions should be discussed against those from other studies, at least in Spain and around the Mediterranean, where similar climatic conditions prevail. 

Figures 3 and4 should be definitely improved. Reductions should be provided as % and not in absolute terms. Also, the colored scale between the right and left columns should be the same in order the reader to be able to follow these results. Furthermore, it’s not clear what the term “marginal” does mean here, and why authors also include the left column “TR”, which is not fully discussed in the main text.

Discussion section should be re-arranged avoiding so many paragraphs, even containing a single sentence!

Furthermore, the impact of meteorology, which seemed to be significant, at least for the precipitation, is not analytically discussed in the part concerning the reductions in air pollutants. It would be nice authors include time series of air pollutants averaged at stations in Madrid center and/or in the peripheral and compare them with the time series in the previous years. This has been followed by several previous studies dealing with evaluation of the COVID-19 lockdown impact on air quality with respect to previous years.   

 

Line 69. Correct as “stations”.

Line 105. Paris is larger than Berlin and also from Madrid.

Lines 139-141. These sentences need revision. The meteorological variables used in this study should be mentioned above and this should be clarified.

169. …on these days…

194-195. This period cannot be considered as pre-lockdown, since it refers to previous years. In general, the whole methodology part is not very well described and the main text needs to be more comprehensive on this issue. Especially the section “pollution maps” includes several mathematical symbols without any equation and a clear determination of the variables.

Line 449. Correct the chemical equation.

Lines 479-482. The effect of meteorology should be more discussed in the manuscript. For example, increased precipitation in 2020 may affect the pollutant concentrations by reducing them. So, the time series analysis has importance, since the effect of precipitation can be defined on them. At any case, this concluding remark gives the sense that the study is not complete at all, and constitutes a main drawback of the paper.