Examining the Economic and Environmental Impacts of COVID-19 Using Earth Observation Data

Round 1

Reviewer 1 Report

Examining the Economic and Environmental Impacts 2 of COVID-19 Using Earth Observation Data

by William Straka III, Bandana Kar, Shobha Kondragunta, Zigang Wei, Hai Zhang, Steven D. 4 Miller, Alexander Watts

General comments:

The manuscript is interesting and may be published but after important  modifications are done, as shown below:

In general:

-All figures except histograms must be redrawn because the text, which is important, is not readable. Following, it is impossible for me to assess whether the associated text, interpretation, suggestions and conclusions are correct or not.

-The many plots (a, b, c, d, e etc) must be “collected” in one figure as a whole. Plots which are part of the same figure are displayed on two separate pages, which again makes the reading very difficult.

-When discussing NO2 decreases in connection with mobility changes, one must look at (relatively)  similar plots for each town. The plots of mobility and plots of NO2 distribution are very different, with completely different scales and one needs a lot of patience to go to and fro. For instance (L385 and around): it is not at all evident to the reader where the commuting areas (probably shown in the first set of figures) are in the NO2 plots in figures 8, 10, 12, to be able to asses that indeed hotspots or cold spots change or do not change.  One could draw some contours of spots in NO2 maps

-What is the square in Figure 2a?

-Chapter 4.3: the decrease could be also due to increasing temperatures; it is well known that NO2 content during cold seasons is higher compared to warm seasons. No temperature variation is shown.

-Lines 363: more argument is necessary in a quantitative evaluation is wanted. I would go for a qualitative discussion, since it is anyway impossible to assess in a credible manner any percentage. Comparing different years will not solve the problem because weather in march 2019 may have been completely different than in March 2020. Thus I agree that some decerase is due to lockdown, but  the discussion must be “relaxed” a bit.  

-L387: Figure 11 shows approximately the same absolute levels NO2 in all months.

-Figure 13: is here Chicago again?

-Fig 14: The caption is not correct, there is no spatial distribution here. I see a relative reduction: compared to what? When? What is grey? Are these monthly averages? If yes, where what is the source?

-Figure 15: as above: The text is unclear; I suggest having smaller plots and bigger text.

-Please explain why is Night time light connected with economic activities and not with household? The reduction in April obviously relates to a significant increase in daylight time. What is red and blue (quantitatively) in figures 15-17?

-I do not see any result about spatial autocorrelation, presented in the end of chapter 3.

-All in all I agree that some changes may exist but not as important as they seem to emerge from the paper. The discussion now ignores almost entirely the natural changes between February, March, April, which is another general problem of the manuscript.

 

 

Author Response

Please see attachement

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors of the article "Examining the Economic and Environmental Impacts of COVID-19 Using Earth Observation Data" raise a very curious issue. Both the topic and the information contained in the article are interesting and worth attention. Nevertheless, I think that the article can be improved and expanded to include some issues that the authors did not raise.

General note: the article requires correction of formatting, for example capital letters, indexes by units, uniform indentation and line spacing.

The method of citation is inadequate to MDPI journals (citation using numbering).

The legends placed on the maps are illegible.

Lines 60–63 "During this time frame, only essential businesses (ex. hospitals, gas stations, grocery stores) were operating, which though helped reduce the transmission of the disease, lockdown also reduced mobility and disrupted economic activities, but improved air quality ." This part fits better in the summary than in the introduction. I suggest the authors to think about it.

Lines 78–80 "Several studies have reported that lockdown measures contributed to emission reduction as measured by the concentration of NO₂, Carbon Monoxide (CO), Sulfur Dioxide (SO₂) and PM2.5 (particulate matter in ug/m³ for particles smaller than 2.5 um in median diameter)." Please explain the abbreviation NO₂ consistently and also add the PM10 mentioned in line 88.

Line 93. I think it is worth explaining the concepts of photochemical smog and sulfurous ‘London-type’ smog.

The authors describe different locations without referring to the climatic conditions that prevail in them. A similar issue concerns the lack of reference to the structure of energy generation in the analyzed locations and the structure of the vehicle fleet (taking into account their age, type, type of fuel used). I believe that such information would be valuable from the readers' point of view.

There is no proper discussion in the article with regard to the literature mentioned e.g. in the introduction. Moreover, I believe that the problem presented by the authors can be expanded. At this stage, the authors should indicate what analyzes or research should be performed in order to comprehensively approach the analyzed issue in the future.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

This work provides a self-explanatory visual analytics approach relating shutdown measures with air quality and economic activity, using novel methodologies and reliable databases. I must recognize an adequate and detailed description of the datasets employed.

 

My main observation is that the linkage 1) between shutdown measures and improved air quality and 2) between shutdown measures and reduced economic activity should be examined by using established statistical techniques (correlation analysis, regression analysis). In section 4.3 Analysis of NO2 it is claimed that “the trend in NO2 is correlated to the mobility pattern observed during the lockdown”. A simple opposition of two histograms is used to substantiate this fact, but such approach supports a simple monitoring of the trends. Statistical results as correlations are proved by statistical evidence (e.g. scatterplot, correlation coefficients, correlation matrix).

 

I would recommend to the author(s) to mention and get feedback from the work by Connerton, P.; Vicente de Assunção, J.; Maura de Miranda, R.; Dorothée Slovic, A.; José Pérez-Martínez, P.; Ribeiro, H. Air Quality during COVID-19 in Four Megacities: Lessons and Challenges for Public Health. Int. J. Environ. Res. Public Health 2020, 17, 5067.

I would ask the authors to contrast their approach with the general approach of Connerton et al.: are their research findings in accordance with the ecosystem impacts that Connerton et al. consider? Authors should describe the added value provided in their own methodological contribution.

 

 

Some other points that require attention are:

 

1. The authors should mention in the text and cite the visualization software (in the references section as well). There is only one mention in line 235 while the entire analysis relies on this software.

 

2. The authors should check thoroughly, in order to avoid grammar, syntax or structure/presentation flaws. Do check the numbers of sections. Some of them are numbered wrongly.

 

3. I would suggest for section 2 a more classical title as “Materials and Methods” or “Data and Methodology” instead of “Study Site”.

 

4. In the Introduction section please define the term “satellite derived economic indicators” more extensively.

 

5. In section 2 Variation in Mobility Pattern (line 308) the term “analysis of income data” is mentioned. Also in line 527 “median household incomes” are mentioned. Could the authors be more specific and cite the income data source?

 

6. Authors mention in section 2 (lines 240-245) Moran’s I index which considers the autocorrelation of spatial units. There is no reference of Moran’s I or Local Moran’s I in the analysis of the following sections. Please define the parts of the analysis where this index is applied and how.

 

7. In the entries of table 1 which is the aggregation measure (statistic) presented for NO2 and mobility? Are the entries the average values over each month?

 

8. In figure 14 please implement color legend. LA is not included in the legend.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Review of

Examining the Economic and Environmental Impacts 2 of COVID-19 Using Earth Observation Data

by

William Straka III*, Bandana Kar, Shobha Kondragunta, Zigang Wei, Hai Zhang, Steven D. 4 Miller, Alexander Watts

General comments:

The authors have responded to some of my comments, but I still would not agree to publishing it as long as some figures are still unreadable. The authors explained why they cannot change some things, however I still think that these changes are possible, with some effort. 

L78 – respective (respectfully means something else)

L135-136 – to be checked.

Figures 5, 6, 7: The numbers are still very difficult to be read in most plots and they need to be enlarged to 250% to be able to see anything. However, after enlarging these as much as my screen allows it, I see that the color code for plots a, b, c is completely different for February, March, April. Moreover, colors are attributed to very large ranges that change from plot to plot. Thus any comparison is left to the reader, that should take each area to se what is actually happening. I don’t think that an analysis of changes in mobility can be accurate if ranges are not equalized (at least). I agree that changes in the maximum mobility may be seen, but the other fine details are not supported by the plots. 

I compared some areas in February and March for LA and I see the following

Center: February: Blue (0-11) – March: blue (0-6)

North East: February: Orange (38 – 88) – March: red (21 – 50) –Orange (11 – 19)

North and North west: February:  Green (18-38) – March: orange (13-21) and red (21-50)

Western tongue: February: Green (18-38) – March: orange (13-21) and red (21-50)

Thus  I see no support for the claims of the authors that “the mobility pattern in March following the lockdown is in direct  contrast to the February pattern (Figure 5a)”

In general the authors should refer to East/NE, center, when discussing plots, since readers around the world do not know where the airports, Malibu, Santa Monica, etc and names in the plots are indistinguishable. 

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,
thank you for considering my comments. From my point of view, the changes introduced to the article are sufficient.

Author Response

Please see the attachment.

Reviewer 3 Report

Table 1 does not prove correlation. The term correlation is not substantiated by the analysis. 

You mention "2018 American Community Survey data available". Which datasets?

The discussion with the general approach of Connerton et al. is not adequate. In fact a couple of lines are added.

The Global Moran’s I results are not visible in the revised manuscript.

Author Response

Please see the attachment

Author Response File: Author Response.pdf