Traffic-related Air Pollution

Air pollution can affect lung development and is implicated in the development of emphysema, asthma, and other respiratory diseases, such as chronic obstructive pulmonary disease (COPD) [1]. PM and nitrogen oxide (NO ₂ ) are linked to chronic diseases. Air pollution affects everyone’s health, but certain groups may be harmed more. Almost 9 out of 10 people who live in urban areas worldwide are affected by air pollution [2].

Nitrogen dioxide (NO ₂ ) is an important contributor to air pollution and can adversely affect human health. NO ₂  in the air comes primarily from the burning of fuel and forms from the emissions of cars, trucks, buses, power plants, and off-road equipment, according to the US Environmental Protection Agency.

Satellite data is used in during the COVID-19 era to understand changes in air pollution. The following images were taken over the first three weeks of March 2020 to compare with March 2019. March 2020 shows less NO ₂  over parts of the United States than at the same time last year. A decrease in NO ₂  concentrations has been reported as a result of lockdown measures to reduce the spread of COVID-19.

The satellite data above shows the overall change in air pollution before and after the COVID-19 lockdown but for health-related studies, the ground-level air pollution measurements are preferred. The satellite data is complex with lower temporal resolutions. Therefore, we used NO ₂  measurements from Environmental Protection Agency's (EPA) public database. The following map shows EPA ground-level monitoring stations all over the US.

We wanted to look more closely into spatio-temporal data of NO ₂  of counties in the US where the population is more than 100,000. We collected daily NO ₂  data from  EPA's database  from March to June of 2019 and 2020. Initially, there were 181 counties where EPA monitors are collecting NO ₂  data. We removed counties with more than 30% of missing data, and 174 counties were included in the final analysis. The map below shows the study area with population proportion in counties of the US.

The pollutant is noticeably lower in concentration in 2020 in New York and California than in 2019, where places including New York City, San Fransico, and Los Angeles may have implemented stricter social distancing guidelines.

We applied time series clustering on time series NO ₂  data of each county in 2019 and 2020. The following maps show an interesting pattern; the counties in New York and in California are again in the same cluster may be strict lockdowns reduce the NO ₂  concentrations which force them to be in the same clusters.

To validate clusters, the ArcGIS time-series clustering generates the F-test and p-test statistics. If p-value is smaller than 0.05 then the cluster is significant

• Pseudo F-statistic for the cluster result is 20.050 • Random Seed: 8820 • Number of basis functions in Fourier: 122

• Pseudo F-statistic for the clustering result is 17.260 • Random Seed: 9801 • Number of basis functions in Fourier: 122

• The GIS analysis of traffic-related air pollution (NO ₂ ) in populated counties has revealed an interesting pattern in the data.

• The NO ₂  is reduced by around 16% on an average in 2020 from 2019.

• The highest mean decrease was 77% in St. Joseph County in Indiana, followed by a 75 % mean reduction in Charleston County in South Carolina.

• The further analysis can be performed to find the effect of AirQuality on COVID-19 infection and mortality in these counties.

[1]  EPA NO ₂  

[2]  https://pubmed.ncbi.nlm.nih.gov/26761628/ 

[3]  (DESCARTES LABS )

[4] Data from EPA  https://aqs.epa.gov/aqsweb/documents/data_api.html