Air quality trends in the last decade

From 2010 to 2022

1 Introduction

Air pollution is one of the most common environmental problem in cities around the world (Morawska et al. 2021). Atmospheric pollutants negatively impact human health and can also affect the climate. To know the the level of air pollution, many cities have installed air quality stations to measure pollutant concentration in the air. These measurements allows us to study the current level of pollution, asses atmospheric emissions regulations, and they support atmospheric research, as they help in model evaluation.

The pollutants that affect human health are knows as criteria pollutants, they are frequently regulated by national and state laws. They are Ozone (O₃), Carbon Monoxide (CO), Nitrogen Dioxide (NO₂), and Fine and Gross Particulate Matter (PM_2.5 and PM₁₀, respectively) (Bekbulat et al. 2021).

In this project, we used the data from the World Health Organization air quality dataset (WHO), to study the trends in PM_2.5 and NO₂ concentration from 2010 to 2022. We focus on PM_2.5 and NO₂, as they are considered, together with tropospheric ozone (O₃), the most harmful to human health (Sicard et al. 2023).

We aim to answer the following questions:

• What continents have more coverage of air quality stations?
• Did PM_2.5 and NO₂ decrease over time?
• What continents have reduced more of their pollution?
• What continents worsen their pollution?

2 Methods

The following code chunk presents the packages we used in this project.

library(tidyverse)
library(ggplot2)
library(gt)
library(sf)
library(rnaturalearth)
library(rnaturalearthdata)

2.1 Dataset

We are using the WHO air quality dataset version 6.1. It is an .xlsx that was read into R using the read_excel function. It was cleaned and saved into a .rds file.

This dataset have yearly average of PM₁₀, PM_2.5, and NO₂, for different cities, from different countries, and WHO regions. Table 1 shows the number of countries and cities in the data by WHO region.

readRDS(here::here("data/processed/who_air_quality_ready.rds")) |> 
  group_by(who_region_name) |> 
  summarise(countries = n_distinct(country_name),
            cities = n_distinct(city)) |> 
  gt() |> 
  cols_label(
    who_region_name = html('WHO region name'),
    countries = html('Countries'),
    cities = html('Cities')
  )

Table 1: Sampled countries and cities in WHO dataset

WHO region name	Countries	Cities
Africa	13	59
America	23	832
South-East Asia	9	399
Europe	49	4347
Eastern Mediterranean	14	159
Western Pacific	13	1379
Non-member state	3	7

2.2 Calculating yearly average

To evaluate the trend of pollutant concentration, we need to calculate the yearly average by WHO region. We also calculate a global average called World and create a single data frame.

who_region_year_avg <- readRDS(here::here("data/processed/who_air_quality_ready.rds")) |> 
  group_by(who_region_name, year) |> 
  summarise(
    pm10_mean = mean(pm10_concentration, na.rm = T),
    pm25_mean = mean(pm25_concentration, na.rm = T),
    no2_mean = mean(no2_concentration, na.rm = T),
  )

who_world_avg <- readRDS(here::here("data/processed/who_air_quality_ready.rds")) |> 
  group_by(year) |> 
  summarise(
    pm10_mean = mean(pm10_concentration, na.rm = T),
    pm25_mean = mean(pm25_concentration, na.rm = T),
    no2_mean = mean(no2_concentration, na.rm = T),
  ) |> 
  mutate(
    who_region_name = factor('World')
  ) |> 
  relocate(who_region_name, .before = year)

who_year_avg <- bind_rows(who_region_year_avg, who_world_avg)

3 Results

3.1 Distribution of air quality stations

Figure 1 shows the distribution of air quality stations that measure PM_2.5 in 2018 in the dataset. It is clear that the northern hemisphere has more coverage than the southern hemisphere, also noted in Garland et al. (2024).

world <- ne_countries(scale = "medium", returnclass = "sf")

readRDS(here::here("data/processed/who_air_quality_ready.rds")) |> 
  filter(year == 2018) |> 
  st_as_sf(coords = c("longitude","latitude")) |> 
  st_set_crs(4326) |> 
  ggplot() +
    geom_sf(data=world) +
    geom_sf(
      pch = 21,
      aes(size = pm25_concentration, fill=who_region_name),
      col = "grey20") +
   scale_fill_manual(
    values=c(
      'Africa'="#79ADE6",
      'America' = "#E67879",
      'South-East Asia'="#9FE778",
      'Europe'="#E179E7",
      'Eastern Mediterranean'="#E8E962",
      'Western Pacific' ="#7AE6CF",
      'Non-member state' ="#FF9933"
      )) +
  labs(
    size = expression('PM'[2.5] * ' (' * mu * '/m'^3 * ')'),
    fill = 'WHO region'
  ) + 
  guides(fill = guide_legend(order = 1, nrow = 4),
         size = guide_legend(order = 2, nrow = 3)) +
  theme_bw(
    base_size = 12
  ) +
  theme(
    legend.title.position = "top",
    legend.position = 'bottom'
  )

Figure 1: Location of cities measuring PM2.5 during 2018

3.2 PM_2.5

Stands for particulate matter with aerodynamic diameter less than 2.5 \(\mu\)m . It is also known as fine particulate matter. They are produce by direct emissions (vehicular emissions or mechanical abrasion) or by inorganic and organic heterogeneous reactions from gas to particle. It is of great interest, as it can impact human health as it can enter to the respiratory system and agraviate cardiovascular and respiratory diseases (Oke et al. 2017) .

Figure 2 shows the variation of yearly concentration in different WHO regions. We can see that Europe and America present the lower concentration over the years. On the other hands, during 2010’s Western Pacific and South East Asia presented higher values, but the later present a continuous reduction of PM_2.5 along the years. Since 2020, there was an increased in Europe and the Eastern Mediterranean regions.

who_year_avg |> 
  ggplot(mapping = aes(
    x = year,
    y = pm25_mean,
    group = who_region_name
  )) + 
  geom_line(aes(color = who_region_name), size = 1.25) + 
  geom_point(aes(color = who_region_name), size = 3) +
  scale_x_continuous(breaks = scales::pretty_breaks()) +
  scale_colour_manual(
    values=c(
      'Africa'="#79ADE6",
      'America' = "#E67879",
      'South-East Asia'="#9FE778",
      'Europe'="#E179E7",
      'Eastern Mediterranean'="#E8E962",
      'Western Pacific' ="#7AE6CF",
      'Non-member state' ="#FF9933",
      'World' = 'black'
      ))+
  labs(
    x = '',
    y = expression('PM'[2.5] * ' concentration ('* mu *'g/m' ^3 * ')'),
    color = ''
  ) + 
  theme_bw(
    base_size = 14
  ) +
  theme(legend.position = 'bottom')

Figure 2: Yearly variation of PM2.5 concentration through WHO regions

3.3 NO₂

NO₂ stands for Nitrogen dioxide. It is gas mainly produce by fuel combustion of vehicules or power plants (Oke et al. 2017). It can cause or agraviate respiratory deseases. It is one of the precursors, together with volatile organic compounds (VOCs), to for tropospheric ozone (O₃).

Figure 3 shows the variation of yearly concentration. Eastern Mediterranean region presented higher concentration of NO2 while cities from Non-member state presented less concentration. There is a global trend in the reduction of NO2.

who_year_avg |> 
  ggplot(mapping = aes(
    x = year,
    y = no2_mean,
    group = who_region_name
  )) + 
  geom_line(aes(color = who_region_name), size = 1.25) + 
  geom_point(aes(color = who_region_name), size = 3 ) +
  scale_x_continuous(breaks = scales::pretty_breaks()) +
  scale_colour_manual(
    values=c(
      'Africa'="#79ADE6",
      'America' = "#E67879",
      'South-East Asia'="#9FE778",
      'Europe'="#E179E7",
      'Eastern Mediterranean'="#E8E962",
      'Western Pacific' ="#7AE6CF",
      'Non-member state' ="#FF9933",
      'World' = 'black'
      ))+
  labs(
    x = '',
    y = expression('NO'[2] * ' concentration ('* mu *'g/m' ^3 * ')'),
    color = ''
  ) + 
  theme_bw(
    base_size = 14
  )  +
  theme(legend.position = 'bottom')

Figure 3: Yearly variation of NO2 concentration through WHO regions

4 Conclusions

• Most of the cities presented in the dataset are located in the Northern Hemisphere. There is an misrepresentation of the Global South, therefore these results are not definitive.
• Europe and America presented the lower values of PM_2.5 along the years, South East Asia, the highest. But, the East Asia region presented a constant reduction along the years.
• The Non-member states presented the lowest values of NO₂ together with America, Eastern-Mediterranean the highest. There is a global reduction of the NO₂ concentrations
• It can be seen a reduction of concentration along the years, but in the case of PM_2.5, there is a minor increment since 2020. For NO₂, the Wester pacific region present an increment since 2019.

5 References

Bekbulat, Bujin, Joshua S. Apte, Dylan B. Millet, Allen L. Robinson, Kelley C. Wells, Albert A. Presto, and Julian D. Marshall. 2021. “Changes in Criteria Air Pollution Levels in the US Before, During, and After Covid-19 Stay-at-Home Orders: Evidence from Regulatory Monitors.” Science of The Total Environment 769 (May): 144693. https://doi.org/10.1016/j.scitotenv.2020.144693.

Garland, Rebecca M., Katye E. Altieri, Laura Dawidowski, Laura Gallardo, Aderiana Mbandi, Nestor Y. Rojas, and N’datchoh E. Touré. 2024. “Opinion: Strengthening Research in the Global South Atmospheric Science Opportunities in South America and Africa.” Atmospheric Chemistry and Physics 24 (10): 5757–64. https://doi.org/10.5194/acp-24-5757-2024.

Morawska, Lidia, Tong Zhu, Nairui Liu, Mehdi Amouei Torkmahalleh, Maria de Fatima Andrade, Benjamin Barratt, Parya Broomandi, et al. 2021. “The State of Science on Severe Air Pollution Episodes: Quantitative and Qualitative Analysis.” Environment International 156 (November): 106732. https://doi.org/10.1016/j.envint.2021.106732.

Oke, T. R., G. Mills, A. Christen, and J. A. Voogt. 2017. “Air Pollution.” In Urban Climates, 294–331. Cambridge University Press.

Sicard, Pierre, Evgenios Agathokleous, Susan C. Anenberg, Alessandra De Marco, Elena Paoletti, and Vicent Calatayud. 2023. “Trends in Urban Air Pollution over the Last Two Decades: A Global Perspective.” Science of The Total Environment 858 (February): 160064. https://doi.org/10.1016/j.scitotenv.2022.160064.