Urban Air Quality Measurements: A Survey

Abstract

Urban air quality is increasingly becoming a cause for concern for the health of the human population. The poor air quality is already wreaking havoc in major cities of the world, where serious health issues and reduction of average human life by a factor of years are reported. The air quality in developing countries can become worse as they undergo development. The urban air quality varies non-linearly depending upon the various factors such as land use, industrialization, waste disposal, traffic volume, etc. To address this problem, it is necessary to look at the plethora of available literature from multiple perspectives such as types and sources of pollutants, meteorology, urban mobility, urban planning and development, health care, economics, etc. In this paper, we provide a comprehensive survey of the state-of-the-art in urban air quality. We first review the fundamental background on air quality and present the emerging landscape of urban air quality. We then explore the available literature from multiple urban air quality measurement projects and provides the insights uncovered in them. We then take a look at the sources that are significantly contributing to polluting the air quality. Finally, we highlight open issues and research challenges in dealing with urban air pollution.

Full text

1
Urban Air Quality Measurements: A Survey
Muhammad Usama∗, Abdur Rahman∗, Zubair Khalid∗, Muhammad Tahir∗, Momin Uppal∗
∗Lahore University of Management Sciences (LUMS), Pakistan.
Email: ∗(muhammadusama, abdur.rahman, zubair.khalid, tahir, momin.uppal)@lums.edu.pk
Abstract
—Urban air quality is increasingly becoming a cause for concern for the health of the human population. The poor air quality is
already wreaking havoc in major cities of the world, where serious health issues and reduction of average human life by a factor of years
are reported. The air quality in developing countries can become worse as they undergo development. The urban air quality varies
non-linearly depending upon the various factors such as land use, industrialization, waste disposal, traffic volume, etc. To address this
problem, it is necessary to look at the plethora of available literature from multiple perspectives such as types and sources of pollutants,
meteorology, urban mobility, urban planning and development, health care, economics, etc. In this paper, we provide a comprehensive
survey of the state-of-the-art in urban air quality. We first review the fundamental background on air quality and present the emerging
landscape of urban air quality. We then explore the available literature from multiple urban air quality measurement projects and provides
the insights uncovered in them. We then take a look at the sources that are significantly contributing to polluting the air quality. Finally, we
highlight open issues and research challenges in dealing with urban air pollution.
F
1 INTRODUCTION
Air pollution is defined as the release of pollutants in the
air that has detrimental consequences on human health
and the planet as a whole. These pollutants can be from
man-made sources or natural sources [1]. Natural sources
of air pollution include fires, sand storms, volcanic activity,
fumaroles, and others. The man-made air pollutants are
gases, droplets, particulate matter, and radiation are emitted
into the atmosphere due to human activity such as burning
wood, coal, gas, oil, alcohol-based fuels, diesel, kerosene,
biomass, waste, etc. It also includes power plants and
chemical factories that emitted toxic gases, particulate matter,
and radiation in the environment. These air pollutants
are causing issues such as acid rains, urban smog, ozone
depletion/holes, indoor air pollution, and global warming
[2]. Air pollution is a complex amalgamation of natural and
human activities. The impact of this relationship is evident
in metropolitan areas (Beijing, Dehli, etc.), where criteria
pollutants, meteorology, infrastructure, and various emission
entities collectively deteriorate the air quality. It is iteratively
reported in the literature that 70 to 80% of the pollution
in the developing world is due to automobile emissions,
where vehicles using low-grade oil on poorly planned road
infrastructure are major contributors to the poor air quality
[3]–[7]. Major cities in the world are suffering from rapid
degradation of the air quality that has pernicious outcomes
on the health of the citizens, economy, plantation, crops, and
livestock [8].
A decline in human life expectancy in metropolitan areas
is accredited to their poor air quality. The problem will get
even worse with the urban development taking place in
underdeveloped countries [9]. In 2013, World Health Organi-
zation (WHO) categorized air pollution as a carcinogen for
human beings [10]. WHO also estimated two million deaths
per year and numerous respiratory illnesses because of poor
urban air quality [11]. The global rise in air pollution has
resulted in a sharp growth in various allergies and respiratory
diseases. The impact of air pollution is not limited to the
metropolitan areas, it also affects the environment on a global
scale, causing health concerns far away from its origin. In
2015, air pollution alone caused 6.4 million death worldwide,
and if the current trend continues, by 2060, the deaths caused
by ambient air pollution will be nearly 9 million people
per year [12]–[15]. In 2015, out of all cardiovascular deaths,
19% were caused by air pollution, similarly, 23% deaths
due to lung cancer were because of air pollution, and air
pollution was the reason for 21% of the total deaths caused
by strokes [15], [16]. Four million new asthma cases and 2
million premature childbirths per year are attributed to fossil
fuel-based air pollutants that cause a dent in the GDP [17].
Furthermore, air pollution appears to be a risk factor (not
yet quantified) in neurodevelopmental disorders in kids and
neurodegenerative illnesses in adults [15], [18], [19].
Air pollution not only affects human health on a global
scale, but it also has an enormous economic cost. The cost
for air pollution emitted by burning fossil fuels in 2018 is
approximately 2.9 trillion USD that is 3.3% of the global
global gross domestic product (GDP) [17]. It is way less than
the money needed to reduce the effect of the air pollution
caused by burning fossil fuels. The toll of air pollution on
the economy is estimated by looking at the six aspects: (1)
cost of human life, (2) people’s ability to work, (3) effects
on the food, (4) reduction in the ability of the ecosystem to
work, (5) damages to the historical monuments, and (6) cost
of remediation and restoration
1
. The economic burden of air
pollution on the GDP of China is 6.6%, for India, it is 5.4%,
for Russia, it is 4.1%, for Germany and US, it ranges from
3.0 to 3.5%, for Japan, United Kingdom, and France it ranges
from 2.0 to 2.5% [14], [20]. It indicates that a monumental
effort is needed to address the air pollution is the need of the
hour.
1. https://unece.org/air-pollution-and-economic-development
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1
© 2022 by the author(s). Distributed under a Creative Commons CC BY license.

2
Figure 1: Organization of the Paper
Another victim of air pollution is agriculture, where bad
air quality is considered a significant contributor to yield
reduction for many decades. Air pollution is rapidly becom-
ing a threat to food production and safety [21]. Effects of air
pollution on human health have been covered rigorously in
the literature compared to agriculture. The adverse effects
of air pollution on the crops vary with the concentration
of the pollutants, geographical locations, and meteorology.
Burning wood and fossil fuel produce sulfur dioxide that
reduces the life and yield of the crops
2
. Rising levels of
acid deposition, ammonia, O
3
, and CO are also affecting the
crops in the developing world
3
. In 2014, India reports a 50%
reduction in the wheat and rice crop yield due to ambient
air pollution [22], [23]. It also highlights the need for serious
reconsideration in environmental policies around the world
to ensure food security.
The global temperature has risen by 1.2
°
C over pre-
industrial levels. The climatic catastrophe is upon us. The
whole world has started feeling the repercussions like
wildfires, heat waves, droughts, etc. Air pollution has played
a vital part in this climatic catastrophe. United Nations
Sustainable Development Goals (UNSDG) 3.9.1 and 11.6.2
directly aim at reducing the mortality rate due to ambient
air pollution and the adverse aspects of particulate air
pollution in urban areas by 2030. Achieving these UNSDG
2.
https://www.britannica.com/technology/agricultural-
technology/The-effects-of-pollution
3.
https://sustainablefoodtrust.org/articles/the-impact-of-air-
pollution-on-crops/
goals for reducing the adverse effects of air pollution in
underdeveloped and developing countries is perhaps a
challenging task.
Identifying the pollution sources, contributions, and
root causes spatiotemporal manner are the vital challenges
associated with urban air quality measurements. Lastly,
based on the spatiotemporal analysis of the urban air quality
making policy recommendations for reducing air pollution
is the motivation for this study. In this paper, we have tried
to answer the following question through a extensive review
of the existing literature:
1)
What are the major air quality modelling and mea-
surement techniques?
2)
What are the major sources of air pollution and how
to best classify them?
3)
What is the situation of the air quality around the
globe and what are the best practices followed for
mitigating the poor air quality?
4)
How the major air quality measurement and im-
provement projects are measuring and dealing with
the urban air pollution and what challenges are
needed to be addressed in order to improve the
effectiveness of these projects?
5)
What are the open research challenges in measuring
the urban air quality?
Contributions of the paper:
In this paper, we build upon the existing literature available
on the air quality measurement and provide a comprehensive
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

3
Paper Survey/review
[24]This paper provides a comprehensive comparison of literature available on static, mobile, and community sensors-based air quality monitoring networks
in the urban environment. It also identifies shortcomings in the existing air quality monitoring networks.
[25]This study examines several environmental sensors and discusses the effects of air pollution on human health. It also gives future guidance in the
development of individual-centric pollution monitoring tools.
[26]Reviewed the low-cost sensor-based system for measuring the air quality and the calibrations of the sensors using machine learning techniques.
The paper also discusses the research challenges and open challenges in using low-cost sensor-based air quality monitoring systems.
[27]Reviewed and summarized the low-cost sensing literature for air quality monitoring. The review also discusses the shortcomings in the data obtained
from the low-cost sensors and open issues in designing low-cost sensor-based air quality networks.
[28]Paper provides a brief survey of the techniques of using chemical sensing, crowdsourcing, IoT, and machine learning in air quality assessment.
Paper provides the results of a two-year air quality monitoring and data collection.
[29] The paper examines the literature on the existing IoT-based low-cost air quality monitoring systems and briefly discusses a few challenges.
[30]
This paper reviews the literature on air quality sensor calibration and identifies the origins of biases and errors in a low-cost air quality sensing network.
It also studies and compares multiple re-calibration techniques of low-cost air quality sensor networks. Lastly, it also provides the limitations and future
avenues in the calibration and re-calibration of the air quality sensors.
[31]The paper conducts a literature review on the low-cost high spatial and temporal resolution air quality monitoring network. It also suggests future
research themes.
[32] This paper reviews the IoT-based air quality monitoring networks and briefly discusses the challenges in designing air quality measurement networks.
[33] The paper provides a comparative analysis of machine learning-based urban air quality prediction techniques.
[34] This paper reviews indoor and outdoor air pollution monitoring using wireless sensor networks.
[35]The paper reviews multiple papers, reports, white papers, and various websites on the role of urban computing in air quality management.
It also covers the techniques of incorporating data-driven mitigation strategies opted by different countries.
[36]This paper reviews the literature on multiple effects of the air pollution monitoring strategies used in South Africa. It also discusses the
challenges involved in designing the air pollution networks in the air pollution monitoring network.
[37]The research compares stationary, dynamic, and pollution data analysis methodologies in depth. The methodology, hardware components,
communication mechanism, assessment, and performance of the air quality system are all compared.
[38]A comprehensive survey on the unmanned air vehicle-based air quality measurement techniques for criteria pollutants along with challenges
and open research directions are covered in this paper.
[39] This research reviews the literature on air quality sensor technologies and air quality management systems.
[40]This paper reviews the air quality standards set by various environmental protection organizations in the world. It also gives an
overview of several aspects of low-cost sensing equipment and methodologies.
[41] This paper reviews the literature on IoT-based machine learning-enabled continuous air quality monitoring and prediction literature.
[42] This paper provides a brief survey of air pollution monitoring systems along with some specific measurement strategies.
[43] This paper gives a summary of the problems involved in monitoring urban air quality.
[44] This paper examines the literature on crowdsourcing-based air quality monitoring and identifies possible flaws as well as future research directions.
[45]Based on the existing literature on the development of an air quality monitoring network, this paper provides the nuts and bolts for
designing the next generation of air quality monitoring networks.
Table 1: Various surveys/reviews on various aspects of air quality measurement.
review of the related work. The major contributions of this
paper are as follows:
•
We provide the fundamentals of air quality measure-
ments along with a non-exhaustive summary of the
air pollutants and their potential sources.
•
We present a comprehensive survey of the techniques
for measuring the urban air quality along with several
sensors famously used for measuring the pollutants.
•
We also discuss the previous/ongoing air quality
measurement projects from various entities and also
summarize a few root cause analyses from the liter-
ature for determining the contributors in urban air
pollution.
•
We also highlight the challenges in designing an air
quality measurement network and how the urban con-
text information can help bring more useful insights
in determining and translating the air-quality.
•
Finally, we highlight the open research issues and
future directions in measuring and learning from
urban air quality.
Organization of the Paper:
The rest of the paper is organized as follows: Section 2
provides a primer on the air quality. It also provides a brief
overview of the air quality landscape of the world while
also covering the details of the major air pollutants and their
sources. Section 4discusses the various approaches available
in the literature for designing an air quality measurement
network. This section also provides details of the various
sensors available for measuring particular pollutants in the
air. Lastly, this section also discusses the diversity in the
air quality data, its relationship with the different context
variables, and how to ensure proper pre and post-processing.
Section
??
provides a comprehensive literature review of the
state-of-the-art in urban air quality standards in the world.
This section also covers projects from various organizations
for measuring and analyzing air quality in different parts of
the world. Section 6discusses the challenges in designing
and measuring the urban air quality and also takes a critical
look at the available literature for providing an exhaustive
list of challenges, trade-offs, tussles, and opportunities in
measuring and analyzing the urban air quality. Section
??
discusses the open research issues and future directions. The
paper has been concluded in Section 7.
2 PRIMER ON AI R QUAL ITY
In this section, we discuss the preliminaries of the air
quality. Then, we provide an air quality landscape and major
pollutants. Lastly, this section provides a discussion on the po-
tential sources of air pollution. Before describing the details
of the air pollutants, it is vital to understand the composition
of pollutant-free dry air. Dry air is essentially a combination
of Nitrogen (78%) and Oxygen (21%). The remaining 1%
is a combination of Argon (0.9%) and extremely minute
quantities of Carbon Dioxide, Methane, Hydrogen, Helium,
and others. Water vapor is also a typical, albeit very variable,
component of the atmosphere, ranging from 0.01 to 4% by
volume; in humid conditions, the moisture content of air can
reach 5%.
4
4. https://www.adb.org/projects/51199-001/main
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

4
Table 2: AQI scale used for indexing the real-time pollution and there impact on human health. Unit followed in this table is
µg/m3unless mentioned otherwise.
AQI Pollution
Level
PM2.5
24 Hour
PM10
24 Hour
CO
8 Hour
(mg/m3)
NO2
24 Hour
SO2
24 Hour
NH3
24 Hour
Pb
24 Hour
O3
8 Hour Cautionary Statement Implications on
Human Health
0-50 Good 0-30 0-50 0-1.0 0-40 0-40 0-200 0-0.5 0-50 None No health risk
51-100 Satisfactory 31-60 51-100 1.1-2.0 41-80 41-80 201-400 0.6-1.0 51-100
The extended outdoor activity
must be avoided by children,
adults, and people with respiratory
issues.
The air quality is adequate;
nevertheless, some pollutants
may pose a considerable health
risk to a limited number of
people who are very sensitive
to air pollution.
101-200 Moderate 61-90 101-250 2.1-10 81-180 81-380 401-800 1.1-2.0 101-168
The extended outdoor activity
must be avoided by children,
adults, and people with respiratory
issues.
Members of sensitive groups
may experience health effects.
The general population is not
likely to be affected.
201-300 Poor 91-120 251-350 10.1-17 18-280 381-800 801-1200 2.1-3.0 169-208
People with respiratory diseases
take precautions and avoid
extended outdoor activities;
everyone else should also
limit outdoor activities.
The general population may
begin to experience health
effects; members of sensitive
groups may experience serious
health effects.
301-400 Very Poor 121-250 351-430 17.1-34 281-400 801-1600 1201-1800 3.1-3.5 209-748
People with respiratory diseases
take precautions and avoid
all outdoor activities; everyone
else should also limit outdoor
activities.
Health warnings of emergency
conditions. The entire population
is more likely to be
affected.
400-500 Severe 250+ 430+ 34+ 400+ 1600+ 1800+ 3.5+ 748+ Everyone should avoid all
outdoor activities
Health alert: everyone may
experience more serious
health effects
2.1 Air Pollution
Air pollutants are particles, gases, or droplets emitted in
the environment that exceeds the environment’s capacity of
absorption, dilution, and dissipation. These pollutants are
gases, solid particles, liquid droplets, etc. The effect of these
pollutants at a scale is termed as air pollution
5
. Air pollution
is increasingly becoming a significant contributor in causing
public health (heart and lung disease, respiratory diseases,
etc.) and environmental issues (global warming, acid rains,
reduction in crop yields, depletion of the Ozone layer, etc.)
at a global scale.
2.1.1 Criteria Pollutants
US Environmental protection agency (EPA) divided air
pollutants into the following six categories that provide
sufficient enough information for determining the overall air
quality are known as “criteria pollutants”.
•Carbon Monoxide (CO):
Carbon Monoxide is a gas
emitted into the atmosphere due to the fossil fuel
burning in automotive vehicles. It has no smell or
color. It reduces the Oxygen supply to the body parts,
thus hindering proper functioning. It also causes
headaches, dizziness, heart, and respiratory issues.
•Nitrogen Oxides:
Nitrogen Oxide is a gas emitted
in the atmosphere due to the fossil fuel burning in
vehicles and power plants. It has a smell and reddish-
brown color. It causes coughs, shortness of breath, and
respiratory infections. It is also a major contributor
to acid rain that is very harmful to crops, plants, and
animals.
•Sulfur Dioxide (SO2):
Sulfur Dioxide is a colorless
gas emitted into the air due to oil and coal-burning
power plants and chemical factories. It has a rotten
egg-like smell. It is a contributor to acid rain that is
harmful to crops, plants, and animals. It is also very
harmful to people with respiratory diseases.
•Ozone (O3):
Ozone is not directly emitted in the
atmosphere, it is a byproduct of the reaction between
5. https://www.britannica.com/science/air-pollution.
Nitrogen Oxide and organic compounds under the
sunlight. Nitrogen Dioxide and organic compounds
emissions are due to a wide range of processes such
as coal/oil-burning power plants, factories, trees, etc.
Ozone here must not be confused with the Ozone
layer present in the stratosphere. It is the main
contributor to smog that can lead to respiratory issues
such as Asthama. It also causes ear, nose, and throat
(ENT) issues. Ozone is also harmful to crops and
plants.
•Particulate Matter:
Solid/liquid droplets suspended
in the air called particulate matter. These particles are
inhalable with a width less than 0.1 mm and a size as
small as 0.00005 mm. PM
10
and PM
2.5
are prime
examples of these particles. PM
10
and PM
2.5
are
inhalable particles with a size less than or equal to 10
micrometers and less than or equal to 2.5 micrometers,
respectively. These particulate pollutants cause lungs
and heart issues and are harmful to crops and plants.
•Lead (Pb):
Lead is a toxic metal with many variants.
It is emitted into the environment by automotive
vehicles burning substandard gasoline. The chemical
factories and power plants are also contributors to
emitting this toxic metal into the atmosphere. Lead
causes kidney issues, strokes, and heart failure [46].
2.2 Sources of air pollution
Sources of air pollution are generally divided into four
categories6:
2.2.1 Natural sources
Natural events are the initial sources of air pollution in
the world. These events are also fundamental parts of the
ecosystem and also had an associated planetary cost. Forest
fires, volcanic eruptions, dust storms, decomposing organic
matter, biological processes in the soil, lightning, and sea
spray are a few examples of the natural events degrading
the air quality. The natural events result in creating different
6. https://www.nps.gov/subjects/air/sources.htm
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

5
Table 3: LIST OF ACRONYMS
WHO World Health Organization
GDP Gross domestic product
UNSDG United Nations Sustainable Development Group
EPA Environmental protection agency
CO Carbon Monoxide
SO2Sulfur Dioxide
O3Ozone
ENT Ear, Nose, and Throat
PM Particulate Matter
Pb Lead
ENT Ear, Nose, and Throat
AQI Air Quality Index
AQLI Air Quality Life Index
NCAP National Clean Air Program
NAAQS National Ambient Air Quality Standard
US United States
NCAP National Clean Air Program
NAAQS National Ambient Air Quality Standard
CASE Clean Air and Sustainable Environment Project
AQG Air Quality Guidelines
EEA European Environmental Agency
MODIS Moderate Resolution Imaging Spectroradiometer
GAM Generalized Additive Model
CTM Chemical Transport Model
POI Point of Interest (POI)
AMS American Meteorological Society
ISCST3 Industrial Source Complex Framework
CTDMPLUS Complex Terrain Dispersion Model
OCD Offshore and Coastal Dispersion
CMAQ Community Multiscale Air Quality
CMAQ-DDM CMAQ Decoupled Direct Method
CMAQ-ISAM CMAQ Integrated Source Apportionment Method
CAMx Comprehensive Air quality Model with Extensions
REMSAD Regional Modeling System for Aerosols and Deposition
UAM-V Urban Airshed Model Variable Grid
CMB Chemical Mass Balance
PMF Positive Matrix Factorization
EPA Environmental Protection Agency
FRM Federal Reference Methods
FEM Federal Equivalence Methods
CRF Conditional Random Field
ARMA Auto-Regression-Moving-Average
LR Linear Regression
NN Neural Network
RT Regression Tree
FEP Frequently Evolving Patterns
GC Granger-causality
EPIC Energy Policy Institute at the University of Chicago
GHAir Ghana Urban Air Quality Project
E-SCRAP Educating School ChildRen to tackle Air Pollution
AI Artificial Intelligence
IoT Internet of Things
types of criteria pollutants, volatile organic compounds, and
biological pollutants.
2.2.2 Mobile sources
Mobile sources of air pollution are considered very deadly for
human health. Here “traffic” encompasses cars, buses, trucks,
trains, planes, etc. Mobile sources are also considered one of
the major sources of air pollution. Air pollution is a result
of the vehicles used for commuting people and resources
[47]. The vehicle exhaust, suspended and re-suspended road
dust, brake dust, and tire wear are sources of traffic-related
emissions [48]. Mobile emissions sources result in different
criteria pollutants and volatile organic compounds with
harmful effects on the ecosystem.
2.2.3 Stationary sources
Stationary air pollution sources include power plants, in-
dustrial facilities, oil refineries, industries, sewage treatment,
and so forth. Stationary sources of air pollution are often
known as ”point sources.” The burning of fossil fuels, metal
processing processes, boilers in industries and power plants,
oil refining procedures, solvents, glues, and paint thinners are
all producers of criterion pollutants such as volatile organic
compounds and hazardous pollutants (mercury dioxin, etc.).
2.2.4 Area sources
Air pollution sources such as agricultural areas, fireplaces,
construction processes in cities, heating and cooling units
in the buildings are categorized as area sources of urban
air pollution. The pollutants from area sources result in
particulate matter and other criteria pollutants. Household
emissions also contribute to the degradation of air quality.
Processes like biomass combustion, fossil fuel burning (such
as coal, diesel, kerosene oil, etc.), tobacco smoking, and cen-
tral air conditioning are a few important sources of household
emissions. Household emissions create different criteria and
biological pollutants. Since this paper only considers ambient
air pollution, indoor air pollution sources are out of the scope
of this work. We also want to note here that multiple sources
from diverse surroundings contribute to urban air pollution,
which varies depending on the geographical location of
the pollution sources in the city, wind direction and speed,
humidity and other meteorological conditions, and so on.
Therefore, attributing urban air pollution to a single pollution
source is an inaccurate approach to look at this issue. The
relationship between criteria pollutants and their sources
is provided in table 4. The table is made based on the
information provided by the US EPA7and NPS8.
2.3 Air quality index
The air quality index (AQI) is a metric used for quantifying
and communicating the air quality in a particular location.
AQI suggest the amount of air pollutant in the air over a
specific average interval. These air pollution concentration
values are measured by a sensor or extrapolated from
a simulation/emulation model. The concentration of the
pollutant and time window is used to determine the dose of
the air pollution, and insights from epidemiological research
provide its health impacts. Based on these health impacts,
a color code and a health advisory are issued for a specific
range of the AQI values. The air quality information varies
for different countries based on their air quality standards
and thus their air quality indices. AQI value for a given
pollutant is determined by the following piecewise linear
function (1).
I=IHigh −ILow
CHigh −CLow
(C−CLow) + ILow ,(1)
where
I
is the air quality index,
C
is the concentration of the
pollutant,
CLow
is the concentration breakpoint that is less
than or equal to
C
,
CHigh
is the concentration breakpoint that
is greater than or equal to
C
,
ILow
is the index breakpoint
corresponding to
CLow
, and
IHigh
is the index breakpoint
corresponding to CHigh.
Measurement data for AQI is averaged over one hour,
there are few pollutants such as Ozone
O3
,
P M2.5
, and
P M10
7. https://www.britannica.com/science/air-pollution
8. https://www.nps.gov/subjects/air/sources.htm
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

6
Table 4: Relation between criteria pollutants and pollution sources categories along with their environmental risks.
Criteria Pollutant Pollution Sources Environmental Risks
Carbon Monoxide (CO) Mobile, Stationary, and Natural
pollution sources Smog and Asphyxiation in vertebrates
Nitrogen Oxides (NOx)Mobile and Stationary pollution
sources
Smog, Acid rain, respiratory issues
in vertebrates
Sulfur Dioxide (SO2)Mobile and Stationary pollution
sources
Acid rain and respiratory issues
in vertebrates
Ozone (O3)Mobile, Stationary, and Area
pollution sources
The main contributor of the smog in urban
areas
Particular Matter (PMx)Mobile, Stationary, Natural, and
Area pollution sources
Haze, Acid rain, serious damages to health
and buildings.
Lead (Pb) Mobile and Stationary pollution
sources
Reduction in biodiversity and neurological
issues
where average over multiple hours is needed to compute a
correct AQI value. Table 2provides a detailed description
of different pollution levels of various pollutants for India
along with their health advisory and impacts on human
health associated with it
9
. Different countries have their air
quality policies and thus have different cutoff values10.
3 AIR QUALITY LANDSCAPE
Before proceeding with the discussions of air quality mod-
eling and measurement, it is imperative to examine the
current global air quality landscape by gleaning insights from
various studies on the impact of air pollution and mitigation
initiatives undertaken in various parts of the world. The air
quality life index (AQLI) [49] report released in July 2020
suggests that air pollution was the most prominent risk to
human health before the pandemic (Covid-19) and after it
as well [50]. Many countries are now putting a lot of effort
into designing policies for reducing emissions, albeit the
progress is slow, and many countries are still struggling to
cope with the air quality issue. In this section, we examine the
air quality landscapes (particulate air pollution) of various
countries, as well as air pollution and the policies used by
these countries to address air pollution challenges.
3.1 Asia
3.1.1 China
China is the most populated country in the world. It is home
to 18.47% of the total population of the world. 61% of its
population lives in cities. In 2013, the concentration of the
PM
2.5
in Beijing city was so high that it seemed that the
city will become uninhabitable [51]. At the time, an average
person in the Beijing city was exposed to approximately 91
µg/m3
of PM
2.5
air pollution. It is nine times higher than the
WHO recommended value for PM
2.5
. In January 2014, the
situation got even worse when the PM
2.5
concentration went
35 to 40 times higher than the WHO recommended value,
and the city officials warned people to stay indoors [52]. The
Guardian describes it as “Beijing’s airpocalypse”. Similarly,
in Shanghai, the air pollution went beyond the critical level,
there the recorded PM
2.5
concentration was six times more
than the WHO recommended value.
Given the situation, in 2014, the Chines government
released a national air quality action plan worth 270 Billion
9. https://app.cpcbccr.com/ccr docs/FINAL-REPORT AQI .pdf
10. https://aqicn.org/scale/
USD with the sole purpose of bringing the air pollution
down. The plan has three goals:
1)
Reduce the PM
10
by 10% relative to its value in 2012.
2)
Reduce the PM
2.5
by 25% in Beijing-Tianjin-Hebei,
by 20% in the Pearl river delta, and by 15% in the
Yangtze river delta.
3) Reduce annual PM2.5of Beijing to 60 µg/m3.
The national air quality action plan worked for China,
by 2017, the PM
2.5
concentration in Beijing-Tianjin-Hebei
went down by 36%. In Pearl and Yangtze delta, the air
pollution went down by 27% and 34% respectively. This
success was achieved due to a collaborative effort from
different government entities in reducing the dependency
on coal, controlling car emissions, increasing renewable
energy, enforcing emission policies, reducing steel and plastic
manufacturing, and replacing coal boilers with natural gas
or electric heaters [53]. Though these steps have improved
the air quality in China, the war against air pollution is not
over, as long-term solutions for bringing air pollution down
to the WHO’s recommended values are needed.
3.1.2 India
India is the 2nd most populated country in the world with
17.70% of the population of the world. 35% of the total Indian
population lives in cities. India is also the 2nd most polluted
country in the world. In 2019, the average PM
2.5
value
was 70.3
µg/m3
that is seven times higher than the WHO
recommended value (10 micrograms/cubic meter). Delhi,
Uttar Pradesh, and northern India are the most polluted
areas where air pollution is reducing almost a decade of
life expectancy of the residents [54], [54]. AQLI India fact
sheet [55] also suggests that 40% of the Indian population
are exposed to air pollution levels not observed anywhere. In
2019, the concentration of the PM
2.5
reached an emergency
level (440 µg/m3).
In 2019, India declared war against pollution and an-
nounced a five-year national clean air program (NCAP) with
42 million USD for the first two years [56]. The goal of NCAP
is to bring the air pollution down by 20 to 30% in 102 cities
(which are over the national ambient air quality standard
(NAAQS)) by building institutional capacity in monitoring
and mitigating the air pollution [56]. The potential impact of
NCAP in the coming years is a 25% improvement in the air
quality and an improvement of 2 to 3 years in the total life
expectancy of the general public [57].
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

7
3.1.3 Indonesia
Indonesia is the 4th most populated country in the world
with a 56% urban population. More than 93% of its popula-
tion is exposed to air pollution that is poorer than the WHO’s
air quality standards. Indonesia is also facing wildfire issues.
In 2015 nearly 100000 wildfires were recorded. The average
PM
2.5
concentration in Indonesia is 40
µg/m3
[58]. Jakarta
is the most congested and one the most polluted city in the
world, 31.5% of the PM
2.5
and 70% of PM
10
particles in
Jakarta air pollution are emitted by the automotive vehicles.
Ten coal power plants around the city are also adding to the
particulate pollution by emitting black carbon [58]. In 1998
the air quality in Sumatra and Kalimantan was below the
WHO recommended threshold. In the last 20 years, the air
quality in these cities has gone three times poorer than the
recommended value. This shift is because of illegal peatland
agriculture, deforestation, and wildfires [58].
The Indonesian government has taken initial steps in
overcoming the air quality issue by adopting the Euro 4
fuel, enforcing automotive health monitoring policies, and
developing a peatland restoration agency. Indonesia’s coal-
based energy production has doubled in the last ten years,
and this is due to the trade-off between the economy and
pollution. A lot of collaborative effort is needed to ensure the
better air quality in Indonesia.
3.1.4 Pakistan
Pakistan is the fifth most populated country with one of
the highest population growth rates (2.0%). On the AQLI
pollution ranking, it is ranked 4th in the most polluted
countries. Pakistan has seen a 20% increase in the PM
2.5
concentration in the last two decades [59]. Lahore has the
poorest air quality in Pakistan, where PM2.5 concentration is
six times higher (64
µg/m3
) than the WHO’s recommended
value [59]. If this level of pollution concentration is sustained,
an average person in Lahore will lose approximately 5.3
years of life expectancy. Almost 99% of the total population
is exposed to pollution levels higher than the recommended
WHO air pollution values [57].
Citing this looming threat, the Pakistani government
started enforcing the air pollution regulations for improving
urban air quality. In 2017 following three initiatives are taken
to ensure improvement of the air quality:
•
Stubble burning is a major contributor to air pollu-
tion in Pakistan. The government of Punjab banned
stubble burning and promoted alternative methods
for getting rid of stubble.
•
Emission regulations were enforced on the vehicles,
factories, and brick kilns.
•
For improving the air quality, Pakistan has also
shut down many coal-based power plants for two
months. This measure has improved the air quality
but resulted in many power outages.
Pakistan can improve air quality sustainably by exploiting re-
newable power sources and continuously enforcing emission
regulations.
3.1.5 Bangladesh
Bangladesh is the 8th most populated country in the world
with a 39% urban population. Bangladesh is also the most
polluted country in the world [57], [60]. The air pollution
there is so intense that an average person loses approximately
6.7 years of life expectancy. Nearly 100% of the Bangladesh
population is exposed to air pollution nearly seven times
more than the WHO recommended air pollution concentra-
tion (10
µg/m3
for PM
2.5
). Major sources of air pollution in
Bangladesh are brick kilns, vehicle emissions, cement facto-
ries, unplanned constructions, and steel re-rolling [60]. In
metropolitans like Dhaka, the concentration of the particulate
pollutants (PM
2.5
and PM
10
) stayed manifold higher than
the recommended air pollution concentration values. The
concentration of other air pollutants like inorganic gases is
noted to stay below the recommended values.
Given the dangerous situation of the ambient air quality
in major cities, the Bangladesh government has started imple-
menting various countermeasures to control and mitigates
air pollution. Bangladesh developed 11 fixed continuous air
quality measurement stations in 8 major cities. The stations
are capable of measuring the concentration of various types
of air pollutants. The recorded data from these monitoring
stations helps develop a spatiotemporal map of different air
pollutants that translates into the identification of the air
pollution trends in the country. Data gathered through these
monitoring stations is also used for developing air models
and AQI for public information.
On the policy front, many initiatives are taken to enforce
the emission policies on brick, cement, and related industries
by banning the import of coal with high sulfur content.
Bangladesh’s government is also incentivizing the industry
to move towards renewable and energy-efficient production
procedures. Initiatives like Clean Air and Sustainable En-
vironment Project (CASE) and Grater Dhaka Sustainable
Transport are also working with the brick, cement, and
transport industries to reduce emissions. Strict enforcement
and monitoring are necessary to ensure the improvement in
the ambient air quality, and the Department of Environment
in Bangladesh has started doing that.
3.1.6 Nepal
Nepal is suffering from a grim air pollution problem. Almost
all of its population is living in an air pollution concentration
higher than the WHO recommended values. According to
the AQLI Nepal fact sheet 2019, Nepal is ranked as the third
most polluted country in the world with an average PM
2.5
concentration of 61.2
µg/m3
that is five times higher than the
acceptable concentration value. The average person in Nepal
is expected to lose at least five years’ worth of life expectancy
if the current levels of air pollution persist. The brick kiln,
fuel burning, vehicle emissions, and road dust are primary
contributors to Nepal’s air pollution. Nepal is far behind in
combating the air quality issues that are affecting the health
of its citizens. More details on the air quality about Asian
countries such as South Korea [61], Thailand [62], etc. are
available on [63].
3.2 Europe
Compared to Asia, Europe already has better air quality. The
majority of Europe’s concentration of particulate pollutants
is below the European Union’s air pollution limits (25
µg/m3
)
but over three-quarters of Europe’s population lives in
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

8
regions that do not satisfy the World Health Organization’s
(WHO) stricter recommendation of 10
µg/m3
[64]. The entire
population of Poland, Belarus, Slovakia, the Czech Republic,
Slovenia, Hungary, Lithuania, Armenia, Belgium, Germany,
Moldova, Cyprus, and Ukraine, and the Netherlands and
San Marino are exposed to pollution levels that do not satisfy
WHO guidelines [64]. Warsaw, Po Valley, and Milan are
three severely polluted areas in Europe. If particle pollution
levels matched WHO guidelines, people would gain one
year and two months [63]. Bursa (Turkey’s industrial center)
suffers from severe particle pollution as well. The population
of Bursa will gain one year and one month if the level of
pollutants are reduced to meet WHO guidelines. Large-scale
biomass burning and unfavorable weather conditions are
causing air quality issues in the Northern Fennoscandia
region (Norway, Sweden, Finland, and Russia) [65]. In the
last two decades, Northern Europe has seen a rise in air
pollution due to several large-scale biomass burning episodes
in Eastern Europe causing serious consequences for human
health and local ecosystems.
According to the European Environmental Agency (EEA)
air quality report 2020, 15% of the European population
(data gathered from 30 countries) is exposed to the PM
10
concentration levels more than the EEA limits and 48%
more than the WHO air quality guidelines (AQG) value for
PM
10
pollutants. Almost 50% of the deployed air pollution
station have reported these statistics. According to EEA
standards for PM
2.5
, only 4% of the population is exposed
to PM
2.5
concentrations higher than the EEA standards.
As per the WHO AQG guidelines, 74% of the European
population was exposed to PM
2.5
concentrations higher
than the recommended values (70% AQ monitoring station
reported these statistics.). According to the same air quality
report, 34% of the population in Europe is exposed to Ozone
concentrations higher than the EEA recommended values.
At AQG levels, approximately 99% of the population is
exposed to Ozone levels higher than the AQG recommended
values. 96% of the air quality monitoring stations have
reported Ozone values higher than the AQG recommended
values. Only 4% of the population is exposed to NO
2
levels
higher than the EEA and WHO AQG values. SO
2
is also on
the decline in Europe, only less than 1% of the European
population is exposed to concentrations higher than the EEA
recommended values and 19% if measured at the WHO AQG
values. Due to Covid-19, statistics reported in the EEA air
quality 2020 report are based on numbers from 2018.
Europe is leading the way in the developed world in
introducing legislation and standards for improving air
quality. Over the years, the EU has developed a procedure
for member countries to access their air quality and share
their data with the EEA. EEA has also provided the member
states with ambient air quality values for twelve major air
pollutants
11
. Table 7provides the standard values of air
pollution concentration for the EU. EU has prescribed the
following principles for member states to measure and report
their air quality12:
11. https://ec.europa.eu/environment/air/quality/index.htm
12.
https://eur-lex.europa.eu/summary/chapter/environment.
html?root default=SUM 1 CODED%3D20%2CSUM 2 CODED%
3D2005&locale=en
1)
Each member state will divide its territory into zones.
2)
Measure the air quality in each zone using sensors,
modeling, or an empirical method.
3)
Report the air quality data to the European Commis-
sion accordingly.
4)
Zones where the air quality is poorer than the air
quality standards (table 7) the member state will
provide a plan to address the sources of emission in
the zone and ensure compliance with the limit value
before the date when the limit value formally enter
into force.
5)
The member state will disseminate the AQI value to
the public.
3.3 United States
The United States (US) is the 3rd most populated country in
the world, with 4.25% of the world population living there.
Over 83% of the total population lives in the cities. The US is
a success story when it comes to air pollution mitigation. In
1970, the US introduced the “clean air act” and after that, the
air pollution gone down by 61% [66]. This decay in pollution
has added 1.4 years to the life expectancy of US citizens.
Los Angeles, once known as the smog capital of the world
now reduced air pollution by 59%. Only 7% of the total US
population is exposed to air quality poorer than the WHO
recommended air quality guidelines [66].
3.4 Africa
West and Central Africa have 27 countries with a 605 million
total population. The average air pollution concentration
(PM
2.5
) is around 20
µg/m3
that is twice the WHO recom-
mended values for the PM
2.5
. With current levels of air
pollution, an average person tends to lose approximately 2.1
years of life expectancy. Benin, Congo (Republic of the Congo
and the Democratic Republic of the Congo), Ghana, Nigeria,
and Togo are among the top air polluted countries in the
region. These countries are also ranked among the countries
having the worst air quality in the world. According to the
AQLI air pollution ranking, Nigeria is ranked 6th in the most
polluted country. In few Nigerian cities (Onitsha, Lagos, etc.),
an average person is expected to lose four to six years of
life expectancy. Brazzaville in the Republic of Congo has
the worst concentration of PM
2.5
(41.5
µg/m3
) and resulting
in 2.3 years of reduction in the total life expectancy of an
average person. The Volta region in Ghana is also suffering
from a poor air quality situation, where the air pollution
concentration is four times the WHO AQG values. The air
quality meeting the WHO AQG values will add three years
to the life expectancy of an average person in Ghana. Burning
fossil fuels is the primary reason for air pollution in Central
and West Africa. Coal consumption is expected to increase
exponentially in the coming years.
The African countries have to strike a balance between
economic growth and air pollution. Air quality data gath-
ering and environment preservation policies are still not
designed. Only Cameron has introduced the National Air
Quality standard for particulate pollution. The African coun-
tries need a coordinated effort to control the emissions and
implementation of air quality standards and environmental
preservation policies.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

9
Figure 2: Nuts and bolts of a comprehensive air quality model. The figure is opted and improved from [67].
4 AIR QUALITY MONITORING,MODELLING,AND
MEASUREMENT TECHNIQUES
In this section, we discuss air quality monitoring, modeling,
and measurement techniques. We divide this section into
four major components along with a necessary discussion
and lesson learned subsection. The four major components
are air quality monitoring networks, air quality modeling
techniques, air quality measurement techniques, and air
quality data.
4.1 Air quality monitoring network
An air quality monitoring network is used to acquire con-
sistent, objective, and standardized information regarding a
region’s air quality. This information may include concentra-
tions of target pollutants. It also allows for necessary steps to
be taken in any environmental protection and public health
safety effort. These steps include determination and control
of emission sources and keeping the public informed about
the state of the air quality [68]. Madruga et al. [69] discuss
the air quality monitoring network with the perspective
of public exposure to pollutants. In literature, air quality
network design usually proceeds in two steps: generation
of the fine spatial distribution of pollutants, and based on
that, optimization of the location of the new sensor to add to
the system. Usero et al. [68] describe the establishment of an
air quality monitoring network in Seville, Spain, to monitor
nitrogen dioxide and ozone levels following the European
Union’s ambient air quality assessment legislation. Mofarrah
et al. [70] have used the multiple-criteria method with spatial
correlation to determine the optimal number of air quality
monitoring stations in an air quality monitoring network in
Riyadh, Saudi Arabia.
By far, efforts in establishing air quality monitoring
networks are broadly categorized in the following groups:
4.1.1 Fixed station air quality monitoring
Fixed air quality monitoring stations are the most reliable,
standardized, accurate, and highly expensive method. Fixed
air quality stations require highly trained staff and resources
to manage the measurement and maintenance operation.
Sometimes these costs even exceed the purchase cost of the
station. Thus, most of the fixed air quality monitoring stations
around the globe are installed and operated by government
agencies. In US, 4000 air quality stations are installed by state
environmental agencies
13
. The EEA receives data from 3000,
2500, and 1000 stations for measuring NO
2
, PM
10
, and PM
2.5
respectively14.
Various efforts have been made to create an ideal air
quality monitoring network that can offer comprehensive air
quality measurements. Elkamel et al. [78] use a multiple-cell
approach to create a monthly spatial distribution for pollu-
tants and use it in a heuristic optimization algorithm to iden-
tify the optimal configuration of a monitoring network. Hsieh
et al. [74] use a semi-supervised inference model to predict
air quality of unknown areas and an entropy-minimization
model to predict the best locations for establishing new
stations. Zhu et al. [73] use Bayesian Maximum Entropy with
a multi-objective optimization model to optimize the design
of an air quality monitoring network. Kang et al. [71] derive
an air quality inference model using a higher-order graph
convolution network. They employed a greedy method to
minimize information entropy, which offers a prioritized list
of places for additional air quality measurement stations to
be installed. The air quality measurement node placement
method [71] enhances overall network performance as well
as air quality prediction for a specific urban area of the city.
As cities continue to expand and city dynamics are always
changing, an optimal method to analyze and redistribute
the installed network is required. Hao et al. [79] employs
an atmospheric dispersion model and genetic algorithm to
maximize coverage with minimum overlap. Yu et al. [80]
use satellite observations to assess the representativeness
of installed air quality stations using a stratified sampling
method. Efforts have also been put into developing, and
assessment of low-cost air quality monitoring alternatives
[81]–[83].
4.1.2 Mobile air quality monitoring
Given the price and the nature of the fixed air quality
monitoring stations, there has been a lot of attention to
the development of mobile air quality monitoring stations.
Mobile air quality monitoring stations offer a cost-effective
solution with several promising features such as high-
resolution spatial pollutant mapping and cross-validation
of air quality measurements. Some of the most prominent
13.
https://www.epa.gov/outdoor-air-quality-data/
air-data-basic-information
14.
https://www.concawe.eu/wp-content/uploads/2018/04/EAQ
Trends digital.pdf
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

10
Papers Input data Spatial distribution generation method Location recommendation technique
[71]1 year of data from 17 stations,
met data, POIs, road networks Graph convolutional neural-network Greedy entropy minimization
[72]Fixed station data integrated with data
generated through COPERT III Generative Adversarial Networks KL divergence and
K-Means clustering
[73] 1 year of fixed station and met data Bayesian maximum entropy Multi-objective optimization
[74]14 months of data from 22 fixed
stations, met data, POIs, road networks Affinity-graph based inference model Greedy entropy minimization
[75]Met data with simulated data from
EPA CMAQ and CAMx models
Simulated through CMAQ and
CAMx models
Objective function and
cost minimization
[70] Generated based on traffic composition Industrial Source Complex (ISC) model Multi-objective optimization
[76]Integration of satellite data with
ground station data Mathematical model Multi-objective optimization
[77] 2 years of sampling campaign Spatial inverse distance weighted interpolation Multi-objective optimization
[78] 6 years of met and pollution data - Multi-objective optimization
Table 5: Review of research on air quality network design
research efforts in developing and utilizing mobile air quality
stations are summarized in Table 6.
4.1.3 Satellite based air quality monitoring
The use of satellite-based sensors for the determination of
air quality has been gaining momentum for a long time now.
Li et al. [102] use MODIS (Moderate Resolution Imaging
Spectroradiometer) data along with meteorological factors to
analyze their relationship with ground-based PM
10
stations.
They use a non-linear regression model to predict PM10
forecast. Fowlie et al. [103] analyze the relationship of ground-
based PM10 stations with satellite-based estimates and their
effect on the Environmental Protection Agency’s policies.
Kim et al. [104] discuss the launch of the GEMS satellite
for monitoring air quality. They discuss the techniques of
sensing different air quality parameters through satellites.
Stebel et al. [105] explore the use of existing satellite data to
derive particulate matter estimates and their correlation with
ground-based stations. They have also extended the sensing
algorithm to report more on the air quality parameters.
4.1.4 Integrating satellite data with ground stations
Sullivan et al. [106] evaluate the need for satellites to
cover the gaps in the existing installed fixed station air
quality monitoring network and the impacts it can produce.
Alvarado et al. [107] have done a comprehensive analysis on
the integration of satellite data into a prediction of ground
air quality for low-income countries. They have used two
models to predict ground-level PM
2.5
, namely, Generalized
Additive Model (GAM) and Chemical Transport Model
(CTM). After analyzing the results, they provide further
recommendations on the ability of satellite estimates to
bolster air quality monitoring networks. Li et al. [108] discuss
the integration of a low-cost air quality sensor network
with fixed ground stations and satellite data to enhance
pollution mapping. Their studies have shown that integrating
the three datasets can vastly improve spatial distribution
and resolution. Their system can also perform quite well
under different weather conditions where the satellite remote
sensing data alone tends to be biased.
4.2 Air quality modelling techniques
The environment is a complex reactive system where multi-
ple physical and chemical processes are happening contin-
uously. The air quality measurement at a specific location
and time provides a conditional spatiotemporal snapshot of
the environment. The interpretation of spatiotemporal air
quality information requires a conceptual understanding
of atmospheric dynamics that is not possible without a
sophisticated air quality model. Measurement alone is also
not enough for policymakers to devise an effective plan
to address the looming challenge of air quality. The air
quality models provide necessary mathematical information
for understanding the complex interactions between different
variables affecting air quality. Therefore a combination of air
quality measurement and air quality models can yield real
progress in understanding and solving the air quality issues
in urban centers.
A comprehensive air quality model is supposed to
take into consideration the meteorology, chemical trans-
formations, emission patterns, known source information,
point of interest (POI), and removal processes and provide
spatiotemporal emission fluxes and pollutant concentrations.
It also highlights the relation between the rate of change
in pollution concentrations and the potential sources [67].
Figure 2illustrates the bare minimum inputs and output
of an air quality model. The three most commonly used
air quality modeling approaches are dispersion, photochem-
ical, and receptor modeling. We briefly describe all three
modeling techniques along with their different variants.
Air quality is not a local phenomenon. Understanding the
contribution of different variables in the air quality landscape
is a challenging task. Modeling these contributions is not
possible through classical air quality modeling techniques.
For completeness, we have included the famous classical
air quality modeling techniques. Though these techniques
are not suitable to model the complex relationships between
different contributing variables in the air quality at a scale,
many advanced modeling techniques are designed based on
the insights from these classical techniques.
•Box models:
The box model is the simplest model
for estimating the concentrations of air pollutants.
The box model compares a domain’s airshed to a
rectangular box within which the pollutant’s mass
is entirely contained [109]. It is used for lab-scale air
quality experiments. It is also suitable for modeling
indoor air quality. More information on the box model
is available on [67], [109].
•Gaussian models:
Gaussian models are the most
popular air quality models used in the literature to
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

11
No. Paper Mobile
Platform
Sensing
Platform
Sensing
Parameters
Study
Area
Time of
the study
1 [84] Mixed Customized
hardware
CO, NO2,
O3California, US 06 Weeks
2 [85], [86]Cycle, bike,
bus, train, walk
Customized
hardware
CO, NO2,
O3California, US 04 Weeks
3 [87], [88]Driving, cycling,
jogging
NODE
sensors CO Sydney, Australia 01 Week
4 [89]Driving, cycling,
jogging
NODE
sensors CO Sydney, Australia 01 Week
5 [83], [90], [91]Walking, driving,
cycling
Teco
envboard PMxGermany 24 Hours
6 [92] Walking Customized
hardware
CO2,
O3Switzerland 06 Month
7 [93]Walking, cycling,
bike, car, bus, train,
Customized
hardware
CO, NO2,
O3California, US 01 Month
8 [94] Cycle, Bike, Car HazeWatch
node
CO, NO2,
O3
New South Wales,
Australia 01 Week
9 [95] Cycle Magee microAeth AE5,
low-cost sensors
PMx, TSP,
Black Carbon, CO Antwerp, Belgium 10 Days
10 [96] Car Custom hardware,
NODE sensors CO, PMx
New York,
New Jersey, US -
11 [97] Bus Customized
hardware PM2.5Hangzhou, China -
12 [98]Google Street
View vehicle
Laboratory grade
analyzers Black Carbon, NOxOakland, US 1 Year
13 [99] Mixed Customized
hardware CO, CO2, CH4India -
14 [100] Trash Trucks Customized
hardware PMxCambridge, US 04 Months
15 [101] Car Customized
hardware CO, CO2, NO Chennai, India -
Table 6: Review of research on mobile air quality sensing
Figure 3: A non-exhaustive taxonomy of air quality modelling techniques from US EPA.
model the repercussions of air pollution in various use
cases. These models are frequently used in regulatory
applications. The Gaussian model assumes that the
plume spread is a result of molecular diffusion. Pollu-
tant concentrations in the plume spread horizontally
and vertically [110]. The solution to the diffusion
equation with varying initial value and boundary
conditions results in a Gaussian distribution of the
pollutant concentrations [67]. For further details on
the Gaussian air quality models we refer the reader
to [67], [110]–[112].
•Eulerian models:
The Eulerian air quality modeling
approach is considered one of the most significant
modeling techniques. It is often known as the “grid
model” technique. In this technique, the area under
consideration is divided into equal size small grid
cells, and conservation of the mass equation is solved
for a specific type of pollutant’s concentrations [113].
A set of mathematical equations in a given coordinate
system explain the transport, diffusion, transforma-
tion, and deposition of pollutant emissions in each cell
[114]. This modeling approach is used for studying
and simulating long-range transport, air quality over
the entire airshed. Further details on the Eulerian air
quality modeling we refer the reader to [67], [113],
[114].
•Lagrangian models:
The Lagrangian models calculate
the wind trajectories and the transportation of the
plume along these trajectories. For source-oriented
models, these trajectories are calculated forward in
time, and for receptor-based models, these trajectories
are calculated back in time [114]. Lagrangian model-
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

12
ing is frequently used to span a longer time duration,
up to years. These models are also used to model the
concentrations of the particulate matter in the air. For
more information on the Lagrangian models, we refer
the reader to [67], [113]–[115].
4.2.1 Dispersion modelling
Air dispersion models formulate and simulate the dispersion
of the pollutants emitted by different sources. The simulation
provides an estimation of the downward air pollutant
concentration. Dispersion models are used to predict the
concentration for specific scenarios such as the change of the
pollution source. These models are more suited for pollutants
that react in the environment and spread over large distances.
The models are also widely used by regulatory bodies during
the preparation and evaluation of air permit applications.
Public safety and emergency response personals utilize these
models to determine toxicity in the air due to possible gas
release events. Environmental protection agencies around the
globe measure the effect of emissions from different sources
and pollution control strategies by using the dispersion
models.
1) AERMOD modeling system:
The American Meteo-
rological Society (AMS)/US EPA Regulatory Model
Improvement Committee (AERMIC), a joint working
group of scientists from the AMS and the EPA,
created the AERMOD dispersion modeling system. It
is a steady-state Gaussian plume model that includes
air dispersion based on planetary boundary layer
turbulence structure and scaling ideas, as well as
handling of complex terrain, simple and intricate
topography. It generates pollutant concentrations
in the ambient air on a daily, monthly, and yearly
basis. It is an updated version of the Industrial
Source Complex (ISCST3) framework proposed by
the USEPA for analysing the influence of industrial
sources on air quality in the coming years [116],
[117]. AERMOD consists of a dispersion model
for short-range dispersion of air pollutants from
various sources, a meteorological data pre-processor
(AERMET), and a terrain pre-processor (AERMAP)
[118], [119]. The AERMOD dispersion model takes
pre-processed meteorological parameters and pre-
processed relation between complex terrain features
and air pollution plumes to produce an air quality
model [117], [120]. Further details on various ver-
sions of its source codes, implementation details,
and variable details are available at [121].
2) CTDMPLUS:
Perry et al. [122] proposed the techni-
cal formulation of the Complex Terrain Dispersion
Model (CTDMPLUS). Later, Paumier et al. [123]
provided a performance characterization study of
CTDMPLUS. The objective for this project by US EPA
was to design a dispersion model that can model
and predict the air pollution concentration in the
mountainous terrain. CTDMPLUS is a point source
Gaussian air quality model for complex terrain that
uses a flow algorithm to provide the deformation in
the plume trajectory caused by the mountainous
terrain. It is capable of simulating the flow and
distortions in the plume near predefined three-
dimensional (3D) terrain features while remaining
simple by applying flow-distortion adjustments to
flat-terrain, Gaussian, and bi-Gaussian pollution
distributions [121], [122]. The CTDMPLUS requires a
significant amount of information on the topography
and weather to produce an efficient dispersion
model, which often represents a bottleneck in many
circumstances. More information on various versions
of the CTDMPLUS source code, implementation
details, and variable details are available at [121],
[122].
3) OCD:
Hanna et al. [124] introduced the Offshore
and Coastal Dispersion (OCD) model, which can
simulate the impacts of offshore emission sources
on coastal air quality. It is based on a steady-state
Gaussian model that can cater to the varying disper-
sion characteristics between over and underwater,
sea-land interface, and aerodynamic effects. Hourly
meteorological data from water and land sites is
necessary for the OCD model to predict air quality.
Turbulence intensities are also frequently used along
with the hourly meteorological data but they are not
mandatory. More information on various versions of
the CTDMPLUS source code, implementation details,
and variable details are available at [121], [125].
Though US EEA recommends AERMOD, CTDMPLUS, and
OCD for air quality modeling, they also provide an alter-
native list of models that can be used by the regulatory
applications on a case-by-case justification.
4.2.2 Photochemical modelling
Photochemical air quality models are often used to evaluate
the effectiveness of the control strategies for regulatory anal-
ysis and attainment demonstrations. Photochemical models
can model the air quality at different spatial scales (local,
regional, national, global, etc.). Photochemical air quality
models are also known as photochemical grid models. These
models are used to evaluate the changes in the concentrations
of the criteria pollutants due to the changes in the associated
variables (meteorological conditions, emission sources, etc.).
Similarly, these models are also used for accessing the
sensitivity of the pollutant predictions in different use cases.
Photochemical grid models are also used to evaluate the
performance of pollution control policies. They simulate the
concentration of the air pollutants at a large scale by using
a complex set of mathematical equations to characterize
different atmospheric processes (physical and chemical). Two
types of commonly used photochemical air quality models
are the Lagrangian trajectory model and the Eulerian grid
model. The Lagrangian trajectory model uses the moving
frame of reference for modeling the air quality whereas, the
Eulerian grid model applies fixed 3D geometric models to the
ground to model the air quality of a particular geographical
area.
1) Community Multiscale Air Quality (CMAQ):
CMAQ modeling system is a state-of-the-art pho-
tochemical air quality modeling system. It uses
3D Eulerian modeling system to simulate the ef-
fect of the criteria pollutants in urban-to-regional-
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

13
to-hemispheric scale. CMAQ
15
is developed and
distributed by US EPA as an open-source suite of
air quality modeling programs that can simulate
multiple air pollution use cases and predicts the con-
centrations based on the historical data of the criteria
pollutants. CMAQ is used to simulate and estimate
the performance of EPA missions for understanding
and forecasting air pollution, human exposure to air
pollution, watershed acidification [126], deposition
of nitrogen and sulfur [127], and many other air
pollution-related use cases. Three common types of
CMAQ are:
•
WRF-CMAQ: Wong et al. [128] proposed a
combination of weather research and fore-
casting with the inputs of the CMAQ (e.g.,
aerosol concentration) to introduce the effect
of the chemistry into the weather. This coupled
design (meteorology from CMAQ and the
chemistry from weather research and fore-
casting component) is known as WRF-CMAQ.
Information from CMAQ, such as aerosol
concentration, is transmitted into WRF so that
the chemistry can influence the weather. More
details on the WRF-CMAQ are available in
[128].
•
CMAQ-DDM: CMAQ decoupled direct
method (CMAQ-DDM) offers concentrations
and deposition sensitivity statistics for user-
specified parameters [129]. The motivation for
CMAQ-DDM comes from the desire to mea-
sure the concentrations of the pollutants by
changing one or a few parameters out of many
predefined air quality model parameters [130].
Air quality models usually take emissions as
input and predict their concentrations CMAQ-
DDM provides the ability to the policymakers
to look at the pollution landscape by tweaking
the parameters of interest or emission sources
(e.g., wildfires, vehicles, etc.). More details on
the CMAQ-DDM are available in [129], [131].
•
CMAQ-ISAM: CMAQ Integrated Source Ap-
portionment Method (CMAQ-ISAM) is a vari-
ant of CMAQ that measures the attribution of
the source in the overall value of the pollutant
concentrations predicted/outputted by the air
quality models. For example, identifying the
proportion of the smog created by the stubble
burning in a neighboring city. This can be
achieved by running the CMAQ twice (first
with all emission use cases and second by
removing the source of interest) but this will
be complex and computationally expensive.
CMAQ-ISAM this issue by calculating source
attribution of many sources directly by the
model in one simulation. Simon et al. [132]
used CMAQ-ISAM for characterizing CO and
Nitrogen oxides sources in the Baltimore area.
Kwok et al. [133] used CMAQ-ISAM to un-
derstand the PM
2.5
sources and their effects
15. https://github.com/USEPA/CMAQ
on the air quality. More details on the CMAQ-
ISAM are available in [130], [133].
2) Comprehensive Air quality Model with extensions
(CAMx):
CAMx
16
is another famous air quality
model in photochemical modeling. CAMx modeling
system models the air quality with all criteria pollu-
tants for a large scale (city, state, country, continent
level). It takes emissions, meteorology data, land use,
surface topography, initial and boundary conditions,
and chemistry-related values as input and performs
source attribution, sensitivity, and process analyses.
Estes et al. [134] used CAMx to model the exceptional
air quality events in near real-time in Taxes to
estimate the ozone impact in three use cases (biomass
burning in Mexico, stratospheric ozone intrusion,
anthropogenic emissions in Mexico). Few critical
resources where CAMx based modeling is used for
air quality policymaking are available in [135], [136].
3) Regional Modeling System for Aerosols and Depo-
sition (REMSAD):
REMSAD
17
is another air quality
modeling system that models the particulate, haze,
and other criteria pollutants. It is a regional scale
modeling system that can simulate the physical and
the reactive processes in the environment to show
the effects of the spatiotemporal changes in the air
pollutant concentration on the overall ambient air
quality.
4) Urban Airshed Model Variable Grid (UAM-V):
In
the early 1970s, the most commonly used air mod-
eling system was UAM-V Photochemical Modeling
System
18
. UAM-V was widely used for air quality
studies focused on Ozone. It is a 3D photochemical
grid model that can model the effects of the chemical
and physical processes in the environment on the
concentrations of air pollutants. UAM-V also pro-
vided a spatiotemporal distribution of the emissions
of various air pollutants. UAM-V is outdated and no
longer used for air quality modeling.
4.2.3 Receptor modelling
The third category of the air quality models is called receptor
models. These models are mathematical techniques for
recognizing and quantifying the origins of air pollution at
a particular receptor location [137]. Receptor models are
different from dispersion and photochemical air quality
models. They do not require meteorological, chemical, and
emission data to estimate the participation of the pollution
sources in the air pollution concentrations at the receptor.
The receptor model uses the chemical and inert properties
of gases (SO
2
, CO, etc.) and the particulate matter particles
to determine to contributions of the emission sources in the
pollution concentrations at the receptor [138], [139].
1) Chemical Mass Balance (CMB):
The CMB [140] is a
model for estimating the contribution of the emission
sources to air pollution at the receptor locations.
CMB uses spatial ambient data and information
16. https://www.camx.com/
17. http://remsad.icfconsulting.com/
18. http://uamv.icfconsulting.com/
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

14
about the pollution sources to determine the source
contributions. CMB quantifies the contributions at
the receptor based on the distinct source types rather
than individual emission sources. A drawback of
CMB is its inability to distinguish between emis-
sion sources with the same chemical and physical
properties. More details on the CMB are available in
[140].
2) Unmix:
Unmix
19
model uses a formula-based on a
form of factor analysis to determine the chemical
species in the air and their sources. It does not take
the chemical profile of the pollution sources as input
instead, it generates the chemical profile to estimate
the number of pollution sources, their syntheses, and
their participation in the air pollution at the receptor
location.
3) Positive Matrix Factorization (PMF):
PMF [141] is
another air quality model which takes different
features from sediments, wet deposition, surface
water, ambient air, indoor air, etc., to identify the
species of air pollutants. PMF also determines the
contributions of the pollution sources at the receptor.
The US EPA no longer updates PMF, and it no longer
supports newer operating systems.
4.3 Air quality measuring sensors
The air quality sensors are the most crucial component of any
air quality monitoring network. These sensors are used to
determine the concentration of pollutants in the air. Typically,
these sensors are built with a Lego connection for the data
acquisition card, and data telemetry is accomplished using
WiFi or cellular communication. In practice, data processing
is done on the cloud rather than on the sensor, however there
have been a few situations where data is preprocessed on
the air quality sensors. Specifications of air quality sensors
are broadly given by the following parameters:
1) Accuracy:
This is a measure of how close the read-
ings of sensors would be as compared to the actual
pollutant value.
2) Precision:
This is a measure of how well the sensor
reproduces the same reading. A sensor with low
precision can give different readings at different
times with the same pollutant level.
3) Range and detection limitations:
This is the mea-
sure of range of pollutant concentration that the
sensor is able to detect correctly. Sensor performance
may vary with different concentration of pollutant.
4) Co-pollutant interference
: Cross interference from
other pollutants also affect sensor readings. It is
intended to minimize the co-pollutant interference
when measuring a certain pollutant.
5) Environmental interference:
Sensor performance
may vary under different environmental conditions
such as low and high temperature, humidity, sun-
light etc.
6) Noise:
Noise is the source of inaccuracy in sensor
readings. The effect of noise should be minimized to
produce more precise and accurate sensor readings.
19.
https://www.epa.gov/air-research/
unmix-60-model-environmental-data-analyses
7) Signal drift:
This is the drift in readings which
occurs due to inherent sensor measurement methods
and degradation of sensor’s components. Many
air quality sensors suffer from signal drift. When
selecting a sensor, it should be ensured that the
drift can either be handled or it does not affect the
readings to a significant level.
8) Response time:
Different sensors can produce read-
ings with different minimum time intervals. It should
be ensured that the selected sensor is able to produce
readings with acceptable time intervals.
9) Multi-site measurement performance:
This is an in-
dicator that generalizes the co-pollutant interference
and the environmental interference of sensor.
4.3.1 Measurement Techniques
A summary of the types of sensors and measurement
techniques has been show in figure 4
•Particulate Matter:
Measurement of particulate mat-
ter concentration is broadly categorized in three
methods: gravimetric, microbalance and optical mea-
surements (Whalley et al. [142]). Gravimetric method
is widely used by regulatory and certification au-
thorities. It is based on weight difference of a filter
medium before and after the gas is passed through
the filter. Microbalance method uses the change
in resonant frequency to measure the particulate
matter concentration. By far the most popular choice
for measurement commercial real-time particulate
matter sensors is the optical method. Scattering or
absorption of a light beam is measured to determine
the concentration of particulate matter.
•Gases:
The two major sensor types used for sensing
gaseous pollutants are electrochemical sensor [143]
and metal oxide sensors. Metal oxide sensors typi-
cally require more power to heat up to very high
temperatures to enable significant sensitivity to target
detection gas. That also increases the start up time of
sensor. Electrochemical sensors require less power to
operate and thus allow for fast startup time and cost
savings generated over life time of sensor20.
4.3.2 Sensing Solutions
•OEM Sensors:
OEM sensors comprise of just the
sensing element. Further interfacing and signal con-
ditioning is required to convert the sensor output
to meaningful numbers. These sensors are popular
choice for original equipment manufacturers.
•Sensing Systems:
Sensing systems are built upon the
OEM sensors and the output is provided in digital
format. Sensing systems are available for both indoor
and outdoor air quality monitoring. Apart from
digital output, many sensors are also available with
extra features such as WiFi or cellular connectivity
and mobile application to view the sensor outputs
as well as the resulting air quality index. For many
commercial and indoor air quality measurement
20.
https://www.emerson.com/documents/automation/
white-paper-electrochemical-vs- semiconductor-gas-detection- en-5351906.
pdf
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

15
Figure 4: The figure shows a taxonomy of air quality sensors along with their working principle and detected pollutants.
systems, high accuracy is not required and thus are
free from regulatory requirements. These systems
are widely available in market and their price is
usually below
$
5000. The data available through such
sensors is not accurate and it can sometimes produce
unknown errors. Weather conditions can also affect
such sensing systems.
To mitigate all of the above problems and to provide
a standard measurement system, the government
defines a certain set of regulations to conform to
for the measurement of air quality and pollutants.
Detailed specifications of measurement and reporting
of the concentration of pollutants is often provided
by the country’s environmental protection agency. In
US, Environmental Protection Agency (EPA) provides
Federal Reference Methods (FRM) and Federal Equiv-
alence Methods (FEM) for measurement of pollutants.
FRMs specify the most scientifically sound technique
to report the concentration of pollutant and it becomes
the basis of criteria for evaluation of other measure-
ment methods. FEM provides techniques that are cost-
effective and easier to implement and yet can provide
comparable level of accuracy with the FEMs. Such
systems are very expensive and their price can range
up to
$
40000. They also highly trained technical staff
for its operation and maintenance. Annual operating
expenses may also exceed the system cost. On the
other hand, the data provided through these systems
is highly consistent and accurate in a variety of
weather and environmental conditions [144].
5 AIR QUALITY MEASUREMENT PROJECTS
5.1 Urban-Air
Urban Air is a Microsoft-funded initiative that began in
2012. It is a sub-project of Microsoft’s Urban Computing
Table 7: Air Quality Standards for the European Union
Pollutants Concentration Averaging
Period
Permitted
exceedences
each year
PM2.525µg/m31 Year N/A
SO2350µg/m31 Hour 24
125µg/m324 Hour 3
NO2200µg/m31 Hour 18
40µg/m324 Hour N/A
PM10 50µg/m324 Hour 35
40µg/m31 Year N/A
Lead (Pb) 0.5µg/m31 Year N/A
CO 10mg/m38 Hours mean N/A
Benzene 5µg/m31 Year N/A
O3120µg/m38 Hour mean 25 days average
over 3 years
Arsenic (As) 6ng/m31 year N/A
Cadmium (Cd) 5ng/m31 Year N/A
Nickel (Ni) 20ng/m31 Year N/A
Polycyclic
Aromatic
Hydrocarbons
1ng/m31 Year N/A
[145], which intends to use big data (e.g., traffic flow, human
mobility, and geographical data) to solve key urban issues
such as pollution, transportation congestion, and energy
consumption. The primary goal of Urban Air was to measure,
analyze, forecast, and assist in the improvement of urban air
quality in cities such as Beijing, China. In addition, Urban Air
sought to discover relationships between various air quality
trends to determine the sources of pollution in different
urban locations.
The Urban Air project consists of four steps:
1) Inferring fine-grained air quality,
2) Forecasting air quality at each station,
3)
Optimal deployment of air quality monitoring sta-
tions,
4) Root cause analysis of urban air pollution.
Following is a brief description of each step.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

16
Table 8: Air Quality Standards for the United States
Pollutant Primary
/Secondary Averaging Time Level Form
Carbon
Monoxide Primary 8 hours 9 ppm Not to be exceeded more
than once per year1 hour 35 ppm
Lead Primary and
Secondary
Rolling 3
month average 0.15 µg/m3Not to be exceeded
Nitrogen
Dioxide
Primary 1 hour 100 ppb
98th percentile of 1 hour
daily maximum concentrations
averaged over 3 years
Primary and
Secondary 1 year 53 ppb Annual mean
Ozone Primary and
Secondary 8 hours 0.07 ppm
Annual 4th highest daily maximum
8-hour concentration averaged
over 3 years
PM2.5
Primary 1 year 12 µg/m3Annual mean, averaged over 3 years
Secondary 1 year 15 µg /m3Annual mean averaged over 3 years
Primary and
Secondary 24 hours 35 µg/m398th percentile averaged
over 3 years
PM10 Primary and
Secondary 24 hours 150 µg/m3
Not to be exceed more than once
per year on an average of
3 years
Sulphur
Dioxide
Primary 1 hour 75 ppb
99th percentile of 1 hour daily
maximum concentrations
averaged over 3 years
Secondary 3 hours 0.5 ppm Not to be exceeded more than once
per year
5.1.1 Inferring fine-grained air quality
As a first step, the Urban Air Project sought to infer air
quality in areas where air quality stations were not available
[146]. It is a difficult undertaking since accessible Air Quality
data is sparse and limited. To tackle this challenge, Zheng
et al. [146] gathered context data from a range of additional
sources (meteorology, road networks, traffic flow, PoI, and
human mobility) that have an indirect impact on air quality.
After acquiring the context data, it is fused with sparse
AQI data from known locations. Then spatial and temporal
classifiers are used for inferring the AQI values at unknown
places. The
Spatial classifier
is an Artificial Neural Network
(ANN) that takes a subset of data and tries predicting
AQI for unknown nodes by using Pearson Correlation of
known features between nodes. The
Temporal classifier
takes the time-dependent factors and tries predicting AQI for
unknown nodes using a linear-chain Conditional Random
Field (CRF). The training is performed by iteratively adding
unknown nodes to the set of known nodes that are classified
confidently by the model(s). For inferring AQI value at
some unknown location/grid, features are applied to each
classifier independently. Only those AQI values are reported
where both classifiers have higher confidence. Results of
the experiment are compared with different interpolation
techniques like Linear, Gaussian, Classical Dispersion Model,
Decision tree, CRF, and ANN. The initial step of Urban
Air only inferred
P M10
and
NO2
values for Beijing and
Shanghai. In subsequent work, Zheng et al. [147] combined
the spatiotemporal model with a real-time feature extraction
database to make user-friendly web and mobile applications.
5.1.2 Forecasting air quality at each station
Following the successful inference of AQI values at arbitrary
sites, the next phase in the Urban Air project was to forecast
AQI values at specific station locations [148]. Forecasting AQI
values is critical because it allows policymakers to better un-
derstand air pollution trends and develop preventative and
mitigation policies. The method of estimating the next AQI
value at a certain time granularity based on prior AQI values
is known as AQI forecasting. Various connected aspects,
such as meteorology, wind speeds, temperature, and so on,
might have an impact on the forecasting process. Therefore,
a real-time database is utilized to give meteorological data
(humidity, temperature, and wind speed) as well as weather
forecasts and AQI values for each station site.
The forecasting component of the Urban Air project
gathers data from 2,296 stations in 302 Chinese cities, with
each instance having concentrations of six air pollutants:
NO2
,
SO2
,
O3
,
CO
,
P M2.5
, and
P M10
. Zheng et al. [148]
employed the four predictors listed below to forecast the
AQI value at a station.
•
A temporal predictor is used to anticipate the AQI
value using a linear regression model on the data.
•
A spatial predictor is used to provide the surrounding
context to a neural network, which forecasts an AQI
value based on the context data.
•
A prediction aggregator trains a regression tree to
give various weights to the first two predictions under
different scenarios.
•
Finally, an inflection predictor is employed to simulate
any rapid changes in AQI values and is only utilized
in exceptional cases (rain, etc.).
For the first 1-6 hours, the AQI value after each hour is
predicted, while for the next 7-12h, 13-24h, and 25-48h,
a min-max range of AQI is forecasted. Zheng et al. [148]
performed forecasting for 36 Air Quality Stations in Beijing
and compared their results with techniques like Auto-
Regression-Moving-Average (AMRA), linear regression (LR-
ALL), neural network (ANN-ALL), and regression tree (RT-
ALL).
5.1.3 Deployment of air quality monitoring stations
The placement of air quality sensors in suitable places is
critical for obtaining relevant air quality data. Hsieh et al.
[74] solve the sensor placement problem with a restricted
budget in the Urban Air project. Hsieh et al. [74] used a two-
stage technique to handle this problem. The first phase is to
infer the air quality at an unknown place, and the second
is to pick candidate locations depending on the confidence
of the inference. The city of Beijing has split into 1km x
1km patches for inference, with each patch referred to as a
node. By linking these nodes with known and unknown AQI
values, a network is formed. Some of the linkages are made
with the help of historical, geographical, and environmental
factors. Local meteorology, road network data, and PoIs are
used to provide weight to edges. Once the graph is complete,
an unknown AQI value is calculated using a weighted sum of
AQI values from known modes. Instead of only obtaining a
single value for AQI, a distribution is learnt at each unknown
node. The weights of the graph are modified over numerous
rounds to get a better distribution (minimize entropy loss).
For recommending the locations of the station, first find
the node with the best distribution (or lower entropy loss)
and rank it the last in the list. After assigning a rank to this
node, label it as a known node and again run the inferring
algorithm to find new inference. After this, find the node
with the best distribution and assign it second last. This step
is repeated until all nodes are ranked.
5.1.4 Identification of root cause of air pollution
The placement of air quality stations is motivated by the
need to identify the root causes of air pollution. Urban Air
also seeks to identify the underlying causes of urban air
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

17
pollution. Understanding the root cause of air pollution is
extremely beneficial to policymakers in government and
environmental protection agencies. Because data collection
involves noise, collecting insights for establishing the root
cause of air pollution becomes a difficult job. Zhang et al.
[149] and Nguyen et al. [150] integrated historical data with
Bayesian learning approaches [151], [152] to uncover air
pollution causal pathways. Data used in [149], [150] includes
measurements of six air pollutants:
P M2.5
,
P M10
,
NO2
,
CO
,
O3
,
SO2
, and five meteorological measurements: tempera-
ture, pressure, humidity, wind speed, and wind direction,
which are updated hourly, for three areas: North China,
Yangtze River Delta, and Pearl River Delta. Understanding
the root cause of air pollution is based on pattern mining and
Bayesian learning. The initial stage in pattern mining is to
locate Frequently Evolving Patterns (FEP). It is accomplished
by first identifying patterns that happen often at the station
and then applying a projection to them, as done by PrefixScan
[153]. Then FEPs of neighboring stations are compared,
and possible causative agents for each sensor/station are
retrieved. The pattern mining module decreases the number
of variables, which aids in decreasing computation cost for
the next stage in learning the BN structure (
2O(n2)
in the
worst-case scenario).
The Bayesian learning module of the root cause identifi-
cation pipeline combines the concentration measurements of
each pollutant at the target node with the spatial data from
candidate causers at the multiple time stamps. Following
data integration, initial routes for the “N” most significant
sensor for the target location/station are created. Each of
these paths is then assigned a Granger-causality (GC) score
[154], [155]. Once the score is assigned to the pathways,
the context data is integrated with pathways. Zhang et al.
[149] determine the number of sub-classes using a “hidden
confounding variable,” and then repeatedly optimize the
initial paths by reducing EM loss.
The Urban Air project was a great success as it helped in
reducing air pollution in China. Based on the insights from
this project, the environmental protection agencies and the
Chines government have taken policy-level steps, and the
air pollution in China is under control.
5.2 AQLI project
The Air Quality Life Index (AQLI) is another famous air
quality measurement project by the Energy Policy Institute
at the University of Chicago (EPIC). AQLI project introduced
a new metric for measuring the impact of air quality called
air quality life expectancy. Instead of the conventional AQI
metric, this new metric translates the impact of air pollution
on the life expectancy of a human being.
The AQLI work is based on the PM
2.5
data collected via
satellite monitoring combined with the global population
data obtained from the 2018 Global LandScan Global Popu-
lation Database [156]. The AQLI index is an extension of the
previous work done by Greenstone et al. in understanding
and quantifying the impact of particulate air pollution on
the expectancy of human life [157]. Once both datasets are
collected, a grid-cell-based procedure is used for combining
global population data with the satellite-driven PM
2.5
data.
Loss in life expectancy is calculated for each grid, where
each grid corresponds to a 6km x 6km area on the ground.
The loss in life expectancy due to PM
2.5
is computed based
on the previous work by Ebenstein et al. [157] which shows
that with every 10
µg/m3
of sustained exposure to PM2.5,
life expectancy decreases by 0.98 years. Assuming that life
expectancy varies linearly with PM2.5 exposure, the loss in
life expectancy is multiplied by 0.98 for each incremental
exposure to 10
µg/m3
of PM2.5 beyond the WHO threshold
level (10µg/m3of PM2.5).
The AQLI project has resulted in AQLI Index
21
which
provides a country-level loss in life expectancy based on
the PM
2.5
concentrations. This project provides a thorough
analysis of the air pollution situation in many countries and
also covers the policy level steps taken by different countries
(e.g., China, India) for mitigating the impact of air pollution
on the life expectancy of their citizens. Though this project
only covers the impact of PM
2.5
, the insights, and policy level
suggestions provided in the AQLI reports can help improve
the air quality of any part of the world. This project is also
an excellent example of how to set up an air quality indexing
study for other criteria pollutants.
5.3 AfriqAir
Air pollution is a major problem in Africa, with research
indicating that air pollution causes around 800,000 premature
deaths every year [158]. Unfortunately, there aren’t many
reference-grade air quality monitoring stations in Africa,
therefore it is difficult to interpolate the actual situation of
the air pollution. The AfriqAir project tries to tackle this
issue by developing a continent-scale air quality monitoring
network. AfriqAir is an African air quality monitoring
initiative [159]. The initiative employs a network of both high-
quality and low-cost air quality sensors. By mid-2020, there
are approximately 50 nodes spread throughout 11 African
nations (Ghana, Rwanda, Uganda, Kenya, South Africa,
Democratic republic of Congo, Cote dI’voire, Niger, Congo,
etc.). AfriqAir has the following three goals for improving
the air quality situation in Africa:
1)
Creating the physical infrastructure required to mea-
sure and monitor air quality across the continent. It
entails a mix of high-quality and low-cost air quality
assessment equipment, as well as the necessary
power sources and data telemetry systems.
2)
Local capacity building to use, manage and analyze
the developed physical infrastructure.
3)
Finally, ensure that the physical infrastructure’s data
and insights are accessible and actionable.
The data gathered through these 50 measurement platforms
across Africa is open-sourced in daily, hourly, and 15-minute
granularity and can be readily used for air quality research
22
.
5.4 Ghana Urban Air Quality Project (GHAir)
Urban Air and AQLI projects are being carried out on a
global scale, using cutting-edge technology and techniques
for planning, modeling, monitoring, and extracting insights.
Many underdeveloped countries with limited resources do
21. https://aqli.epic.uchicago.edu/the-index/
22. http://www.afriqair.org/
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

18
not have this opportunity. With the Ghana Urban Air Quality
Initiative (GHAir), we hope to show the reader how a
relatively impoverished nation may set up an air quality
monitoring project.
Ghana is West Africa’s second-most populated country,
behind Nigeria. Ghana is dealing with severe pollution
problems, and it is reported, that air pollution killed over
28,000 people in 2018 [160]. In 2019, Ghana started the GHAir
project to solve the problem through cutting-edge research.
The environmental protection agency of Ghana has limited
resources and proper technical human resources. To solve
these issues, GHAir uses low-cost air quality sensors to
bridge the data gap. Various studies have highlighted that
low-cost sensors might be a great chance to overcome the
air pollution data gap in underdeveloped countries [143],
[161]–[164].
GHAir has the following four objectives that are also in
line with the UNSDGs [165]:
1)
Creating a dense low-cost air quality sensor network
in metropolitan areas to collect real-time spatiotem-
poral air quality data that may be used to impact air
pollution management policy.
2)
Launching public awareness campaigns about the
effects of urban air pollution and how residents may
safeguard their health in areas with poor air quality.
3)
Improving the air quality by introducing behavioral
changes in the communities.
4)
Performing epidemiological research to highlight the
health issues of air pollution exposure in vulnerable
populations for the public health department.
The GHAir presently employs a mix of low-cost Pur-
pleAir sensors, Clarity nodes, RAMPs, and Modulair-PM
sensors. These sensors have been installed in six of Ghana’s
major cities (Accra, Tema, Cape Coast, Takoradi, and Kumasi).
GHAir has also just placed ten sets of TEOM 1400ab PM
monitoring sensors provided by the UK Environmental
Agency (Automatic Urban and Rural Network) [165]. The
GHAir presently employs a mix of low-cost PurpleAir
sensors, Clarity nodes, RAMPs, and Modulair-PM sensors.
These sensors have been installed in six of Ghana’s major
cities (Accra, Tema, Cape Coast, Takoradi, and Kumasi).
GHAir has also just placed ten sets of TEOM 1400ab PM
monitoring sensors provided by the UK Environmental
Agency (Automatic Urban and Rural Network) [165]. GHAir
project is also going to launch a program called E-SCRAP
(Educating School ChildRen to tackle Air Pollution) project
with the help of the Royal Society. The project’s goal is to
create awareness among schoolchildren about air quality and
how they may help to improve it. The motto of the project is
“School children as agents for improved air quality”.
They are now experiencing several difficulties in getting
data from sensors. These concerns include the availability of
WiFi at deployment sites for data telemetry and sensor power
supply. To address these issues, GHAir is experimenting with
solar energy to power the sensors [165]. Furthermore, they
are attempting to leverage GPRS for data telemetry. Despite
these challenges, the GHAir project has enormous potential
for bridging Ghana’s data gap on air quality.
5.5 Hazewatch
PM
2.5
concentrations above WHO standards have been
found in New South Wales, Australia, particularly in Sydney.
The Department of Environment, Climate Change, and Water
(DECCW) has already placed 15 stations in various sites
across Sydney, and data is published hourly. AQI levels
and corresponding health advisories are provided based
on this pollution data. Unfortunately, these stations are
separated by tens of kilometers, resulting in inadequate
spatial resolution. Because of the low spatial resolution,
complicated interpolation procedures are needed to report
AQI values. As a result, the DECCW AQI monitoring
network does not represent actual levels of air pollution and
exposure. To overcome these shortcomings, Sivaraman et al.
[87] designed a low-cost urban air quality monitoring system
known as HazeWatch. HazeWatch utilizes many low-cost
mobile sensor units installed in cars to measure air pollution
concentrations, as well as users’ mobile phones to tag and
upload data in real-time. The outcome of the projects is its
cost-effectiveness, better spatial resolution, and personalized
exposure tools. The project measured NO
2
, CO, and O
3
.
Though HazeWatch filled the gap in spatial resolution, it has
faced multiple challenges in calibration, sensor design, mass
deployment, health outcome interpretation, etc. The project
has resulted in multiple research publications on designing
pollution monitoring sensors [166]–[168], data transmission
[169], [170], database connectivity [171], android interface
design [172], pollution modeling [87], data visualization [173],
and exposure modeling [174].
5.6 CITI-SENSE
Developing a country-wide air quality network based on
reference-grade, near-reference, and low-cost air quality mon-
itoring sensing solutions for each pollutant is a complicated
endeavor. It requires a lot of money, infrastructure, and
technical expertise; thus environmental governance through
citizen empowerment is gaining traction. The purpose of
these programs is to encourage individuals in deploying low-
cost, sometimes near-reference-grade monitoring equipment
and share air quality data. The integration of data from
residents and reference-grade sensing equipment can assist in
obtaining a more granular picture of air quality, resulting in
more effective air quality improvement measures. The Euro-
pean Commission has financed “CITI-SENSE”
23
(2012-2016),
a project that uses cutting-edge Earth observation technology
to build and test environmental monitoring systems. The
project’s goal was to create citizen observatories that would
allow residents to gather and monitor environmental data
in order to formulate community policies. The CITI-SENSE
project produced a number of air quality sensor devices, as
well as mobile and other communication technologies. The
key contributions of the CITI-SENSE project were:
•
Studying and mitigating hurdles in the citizens’ in-
volvement in environmental decision making.
•
Designing the tools and technologies to enable the
citizens in collecting urban environmental data.
•
Providing low-cost measuring solutions and data
fusion methods for scientific analysis.
23. https://cordis.europa.eu/project/id/308524
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

19
•
Integrating newer sensing technologies (IoT and other
ICT technologies), cloud platforms, data analysis, and
learning techniques to enhance community participa-
tion in the form of personal environmental monitoring
devices.
The CITI SENSE initiative distributed 324 air quality mon-
itoring units around Europe, and 400 volunteers helped
test the personal air quality monitoring devices. A total of
24 citizen observatories were also developed in nine major
cities of Europe (Barcelona, Belgrade, Edinburgh, Haifa,
Ljubljana, Oslo, Ostrava, Vienna, and Vitoria-Gasteiz). An
air perception application was also a critical part of the
project, and 1200 people used it to report and get air quality
information. In 2015-2016, nearly 9.4 million environmental
observations were collected using the CITI-SENSE sensor
network and other additional observation tools. Further
involvement of the citizens was ensured by feedback surveys,
questionnaires, focus group discussions, and interviews. The
CITI-SENSE initiative resulted in environmental monitoring
systems across Europe as well as citizen engagement in
environmental governance. The insights from the CITI-
SENSE project has resulted in many research publications
dealing with low-cost air quality sensing and performance
assessments [81], [175]–[178], pollution hotspot detection
[179]–[181], data assimilation [175], [182], missing data
imputation methods [183], epidemiology studies [184], air
quality sensor calibrations [175], [185]–[189], localized real-
time pollution effects [177], [180], [181], [190], zero emission
studies [191], wireless and distribution network design
suggestions for air quality networks [192], pollution exposure
assessment [81], [180], end-user feedback [180], [193], [194],
toolkits for monitoring urban air quality [195], and new
citizen observatory design [196].
5.7 OpenSense II
Generating a comprehensive spatiotemporal map of air
pollution requires a lot of data from multiple sources. Only
reference-grade air quality data is not enough as they are
very expensive and there can be a few reference-grade air
quality monitors in a city. OpenSense II aims to integrate
data from heterogeneous devices and crowdsourcing with
reference grade measurements to generate a spatiotemporal
map of urban air pollution and estimate the health impacts
due to air pollution exposure. OpenSense II project generates
granular air pollution maps of Zurich and Lausanne and also
studies the impact of air pollution exposure on human health.
OpenSense II uses data (air pollution data, communication
platforms, sensors, personalized health recommendations,
etc.) from another project known as “Nano-Tera project
OpenSense”. The data from the Nano-Tera project is com-
bined with crowdsourcing and human-centric computation
techniques for high-resolution air pollution maps. The air
quality data is also gathered by deploying sensing systems
on buses and electric cars. OpenSense II also pushed the
state-of-the-art in generating high-resolution spatiotemporal
air quality maps [
?
], [197], [198], mobile sensor networks
for air quality monitoring [199], [200], and estimating the
impact of PM
x
on human health and personalized health
recommendations [201].
5.8 Root cause analysis of the urban air pollution
The efforts put into designing, modeling, measuring, and
developing cutting-edge air quality measurement facilities
are only to understand the root cause of urban air pollution
and how it affects human health and the global tempera-
ture. Many studies have been conducted using air quality
data acquired from air quality networks to identify the
causes/contributors of air pollution. Karagulian et al. [202]
performed a systematic analysis on the air quality data of 51
countries from the WHO website and highlighted the major
sources of air pollution. According to their study of available
data, traffic emissions contribute 25% of air pollution (PM
2.5
),
industrial activities contribute 15%, domestic fuel burning
contributes 20%, natural dust and salt contribute 18%, and
unidentified causes linked to humans contribute 22% [202].
Jiang et al. [203] investigated the spatiotemporal features of
air pollution in six Chinese cities and applied the Granger
causality test [204] to evaluate the impact of a city’s air
quality on surrounding cities and vice versa. According to
their study, air pollution is very high in the winters and
early springs and stays low in summer and autumn. They
also discovered a unidirectional association between the air
quality of Baoding and Beijing, where the air pollution from
Baoding has a significant impact on Beijing’s air quality
(since Baoding is more polluted than Beijing) [203].
Wang et al. [205] found that particulate matter from
transportation, industry, agricultural activities, fuel burning,
construction, and demolition accounts for 85 to 90% of overall
air pollution in China. Wang et al. [205] also discovered
that the 2013 extended Haze event in central-eastern China
was caused by a shift in meteorological conditions. [205]
employed synthetic atmospheric circulation to determine
the sources of air pollution. Traffic emissions and high
levels of energy consumption are identified as contributors
to the Haze and poor air quality in central-eastern China
[205]. Recently there has been a surge in data-driven root
cause analysis techniques [206]–[209]. These techniques are
motivated by the success of big data and artificial intelligence
(AI) in many other domains. For detecting and comprehend-
ing the root cause of urban air pollution, we advocate a
combination of traditional modeling/causal methodologies
with cutting-edge AI techniques.
6 UR BAN AIR QUA LIT Y: CHALLENGES
Despite a plethora of work in measuring and understanding
air quality, there are various challenging aspects in tracking
and measuring air quality. In this section, we take a look at
the challenges needed to be addressed for rapid improve-
ment in ambient and indoor air pollution.
6.1 Data collection and public datasets
Collecting the air quality data is a challenging task as it
involves different concentrations of air pollutants. Given the
environmental cost and health risks of poor urban air quality,
it is imperative to develop a central real-time air quality
data measurement and processing system. Two paradigms
for gathering urban data (i.e., air quality data, POI, mete-
orological data, etc.) are sensor-centric data collection and
crowd-centric data collection. The sensor-centric paradigm
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

20
has two categories. These categories are based on whether
the data collecting sensors are mobile (deployed on a moving
object) or static (deployed on a fixed location). The crowd-
centric data collection is also divided into two categories;
active(data generated via participatory surveys and check-
ins) or passive (data generated by users passively while using
the urban infrastructure).
Gathering air quality and related data via these two
paradigms is difficult because of the following challenges:
•
In static sensor-centric air quality data collection,
sensors are deployed at a fixed location, and they
communicate at a predefined frequency to the cen-
tral database (i.e., cloud server). This static sensor
deployment makes the air quality measurement a
resource constraint system (limited budget, land use,
workforce for maintaining the system, etc.). As a
result, data gathered from very few sensors fixed at
different locations in an urban environment is sparse
and not sufficient representative of the air quality
situation of the city.
•
In static sensor-centric air quality data collection
where the number of sensors is limited, the optimal
placement of the air quality sensors for gathering
representative enough data becomes a challenge.
•
Though mobile sensor-centric data collection help
resolve the issues faced in air quality data gathering
due to the fixed nature of the static sensor-centric
approach, it has its challenges. The air quality data
gathered from the sensors mounted on moving objects
such as buses, bikes, taxis, UAVs, etc., is skewed by
the movement of these moving objects. For example,
buses are usually used as a means of the commute
from a busy fixed route, the gathered data will
provide a good representation of the air quality along
the bus’s route but it will not provide a true depiction
of urban air quality.
•
Another challenge in mobile sensor-centric air quality
data is the redundancy in the collected data. Since the
sensors-mounted vehicle will be following a route (es-
pecially buses), the data collected will be redundant,
and the data from less traversed routes will be sparse.
It will result in an imbalanced data distribution. Any
deduction made from the distribution will be biased
towards the most frequent route.
•
Human as a sensor is another way of getting the
data for inferring the air quality. Data generated
by the citizens passively while accessing the urban
infrastructure (call data record, public Wifi, passenger
bus card swipes, taxi pick and drop locations, etc.) is
useful in determining the context of urban pollution.
Here the key challenges are the privacy of the users,
the security of the service providers, and meeting the
legal requirements of data protection.
•
Participatory crowdsensing is a procedure opted for
gathering the context data used for inferring the urban
air quality where the measurement campaigns and
surveys are used to collect the data. imbalance data
coverage, unavailability of ground truth to measure
the quality of the collected air quality data, and
noisy and fake data reporting are a few challenges
associated with the participatory crowdsensing for air
quality-related data.
To overcome these challenges in air quality data collection
along with its proper context requires a great deal of planning
and understanding of the urban environment. Selecting the
right kind of sensors, an acceptable level of measurement
granularity, designing a proper measurement campaign,
choosing the appropriate cloud/database, and ensuring the
quality of the collected air pollution data and the motivation
to make it public can help develop a comprehensive air
quality dataset for determining the correct air quality values.
Nevertheless, a hybrid approach combining sensor-centric,
expert-in-the-loop data collection techniques can yield better
air quality data collection.
6.2 Air quality monitoring networks in underdeveloped
countries
The air quality situation in underdeveloped nations is dire,
yet they are unable to address it due to a lack of air quality
monitoring networks. Deploying these networks across the
country necessitates large sums of money and planning that
in many underdeveloped countries is not available. Without
an adequate and extensive air quality monitoring network
in place, economically developing nations are unable to
gather air quality data and, as a result, lack policies for
monitoring and combating air pollution trends. Low-cost air
quality monitoring stations are used instead of actual weather
stations and deployed in a few countries. Developing a
country-wide low-cost air quality sensors-based network
remains a very challenging task.
6.3 Trade-off between economic growth and air pollu-
tion
The trade-off between economic growth and air pollution im-
plies that economic expansion is connected to industrializa-
tion and transportation, which necessitates the combustion of
gasoline and other energy sources, resulting in air pollution.
Finding the right balance between economic growth and
air pollution is a daunting task for many countries. Other
variables such as population, urban density, and urban
planning exacerbate the difficulty of this task.
6.4 Regularization and air quality measurements
Even though there are several environmental regulatory
bodies and a plethora of regulatory rules but we continue to
see that air quality in urban areas is becoming a predicament.
The implementation of these standards is an issue that
many governments are unable to address for a variety of
reasons, including a lack of education, financial resources,
political and religious divisions, lack of global and regional
cooperation, and so on [210].
6.5 Urban planning and air quality
Urban planning plays a vital role in improving the air
quality of the city. Unfortunately, many metropolitans around
the world are suffering from the worst air quality due to
exponential growth in population, traffic congestions, high
built densities, and lack of urban planning. Canyon street (the
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

21
street design in which both sides of the roadway are bordered
by buildings) has a poor dispersion rate, thus vehicle exhaust
remains in greater quantities than usual, producing severe
health issues and air pollution. Avoiding canyon street design
in urban planning might help to reduce urban air pollution.
Reducing urban air pollution from traffic is strongly linked to
initiatives for encouraging active commuting (travel by low
energy-consuming vehicles) and lowering carbon emissions.
Urban planning and its relationship with different urban
environmental concerns is an area in which much ingenuity
is required.
6.6 Personalized context-aware air quality measure-
ment applications
Designing context-aware air quality monitoring systems are
gaining traction. Air quality is heavily reliant on several
context factors (PoIs, meteorology, etc.), and interpreting
air quality measurements without taking the context into
account might create bias in the measurements. People
with respiratory or ocular diseases are especially sensitive
to air pollution and should be warned about it. The AQI
measurements are insufficient for these patients. It is very
challenging to develop context-aware custom-made air qual-
ity monitoring applications. A few emerging applications use
the Internet of Things (IoT) and tailored context to deliver
customized warnings on the severity of air pollution in a
specific city location [29], [211]–[217].
6.7 Impact of climate change on the air quality
Climate change has started causing many problems in
different parts of the world. Climate change can influence the
local air quality and vice versa. An increase in the ground-
level O
3
is observed as the atmosphere gets warmer due
to climate change, and this ground-level O
3
is expected to
cause dense smogs in urban areas. The jury is out on the
effectiveness of climate change on particulate matter-based
air pollution 24.
6.8 Indoor air pollution
Air quality within the buildings (houses, schools, shopping
malls, airports, etc.) concerning the health of the people
health is termed as indoor air quality (IAQ). It is described
in the literature that the IAQ in homes is 2 to 5 times
more polluted than the ambient air pollution [218]. CO,
microbiological contamination owing to moisture, insuffi-
cient ventilation, fuel burning, incorrect building design,
and commonly used construction materials are a few of
the causes of increased indoor air pollution levels. Long-
term (respiratory disorders, cancer, heart disease) and short-
term (ENT irritations, headaches, tiredness, nausea) health
problems can be caused by poor IAQ. People have been
staying indoors for the last two years as a result of the Covid-
19 restriction, and indoor air pollution in homes [219], [220],
and hospitals has skyrocketed [221]. With this exceptional
circumstance, IAQ improvement is both vital and challenging
[222].
24.
https://www.epa.gov/air-research/
air-quality-and-climate-change-research
7 CONCLUSIONS
The urban air quality is turning out to be daunting health and
economic challenge for the metropolitan centers around the
globe. The lack of measuring infrastructure is making this
situation even harder. This paper provides a non-exhaustive
yet comprehensive survey of the urban air quality measuring
methodologies, standards, and initiatives operating through-
out the globe. We have also emphasized the challenges
restricting the urban air quality measurement. We also
invite the readers to ponder upon these challenges and offer
suggestions for better air quality in our cities.
REFERENCES
[1]
Wayne T Davis, Joshua S Fu, and Thad Godish.
Air Quality
. CRC
Press, 2021.
[2]
Mark Z Jacobson.
Air pollution and global warming: history,
science, and solutions. Cambridge University Press, 2012.
[3]
Auto Fuel Policy. Report of the expert committee on auto fuel
policy–executive summary.
Ministry of Petroleum and Natural
Gases, Government of India, 48, 2002.
[4]
Haikun Wang, Lixin Fu, Yu Zhou, Xuan Du, and Weihua Ge.
Trends in vehicular emissions in china’s mega cities from 1995 to
2005. Environmental Pollution, 158(2):394–400, 2010.
[5]
Ye Wu, Shaojun Zhang, Jiming Hao, Huan Liu, Xiaomeng Wu,
Jingnan Hu, Michael P Walsh, Timothy J Wallington, K Max
Zhang, and Svetlana Stevanovic. On-road vehicle emissions and
their control in china: A review and outlook.
Science of the Total
Environment, 574:332–349, 2017.
[6]
Shida Sun, Luna Sun, Geng Liu, Chao Zou, Yanan Wang, Lin Wu,
and Hongjun Mao. Developing a vehicle emission inventory with
high temporal-spatial resolution in tianjin, china.
Science of The
Total Environment, 776:145873, 2021.
[7]
Shida Sun, Jiaxin Jin, Men Xia, Yiming Liu, Meng Gao, Chao Zou,
Ting Wang, Yingchao Lin, Lin Wu, Hongjun Mao, et al. Vehicle
emissions in a middle-sized city of china: Current status and
future trends. Environment international, 137:105514, 2020.
[8]
Jes Fenger. Urban air quality.
Atmospheric environment
,
33(29):4877–4900, 1999.
[9]
Hannah Ritchie and Max Roser. Urbanization.
Our World in Data
,
2018. https://ourworldindata.org/urbanization.
[10]
World Health Organization (WHO) et al. Health risks of air
pollution in europe.
HRAPIE Project. Copenhagen, Denmark:
WHO, 2013.
[11]
World Health Organization et al. Who. air quality guidelines for
particulate matter, ozone, nitrogen dioxide and sulphur dioxide.
global update 2005.
World Health Organization. Available from:
http://www. euro. who. int/ data/assets/pdf file/0005/786
,
38:E90038, 2006.
[12]
GBD 2015 Risk Factors Collaborators et al. Global, regional,
and national comparative risk assessment of 79 behavioural,
environmental and occupational, and metabolic risks or clusters
of risks, 1990–2015: a systematic analysis for the global burden of
disease study 2015.
Lancet (London, England)
, 388(10053):1659,
2016.
[13]
World Health Organization et al. Ambient air pollution: A global
assessment of exposure and burden of disease. 2016.
[14]
E Lanzi et al. The economic consequences of
outdoor air pollution.
Organization for Economic
Cooperation and Development. Retrieved from
https://www.oecd.org/environment/indicators-modelling-outlooks/Policy-Highlights-Economic-consequences-of-outdoor-air-pollution- web.pdf
,
2016.
[15]
Philip J Landrigan. Air pollution and health.
The Lancet Public
Health, 2(1):e4–e5, 2017.
[16]
Haidong Wang, Mohsen Naghavi, Christine Allen, Ryan M
Barber, Zulfiqar A Bhutta, Austin Carter, Daniel C Casey, Fiona J
Charlson, Alan Zian Chen, Matthew M Coates, et al. Global,
regional, and national life expectancy, all-cause mortality, and
cause-specific mortality for 249 causes of death, 1980–2015: a
systematic analysis for the global burden of disease study 2015.
The lancet, 388(10053):1459–1544, 2016.
[17]
Lauri Myllyvirta. Quantifying the economic costs of air pollution
from fossil fuels.
Centre for Research on Energy and Clean Air
,
2020.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

22
[18]
Philippe Grandjean and Philip J Landrigan. Neurobehavioural
effects of developmental toxicity.
The lancet neurology
, 13(3):330–
338, 2014.
[19]
Marianthi-Anna Kioumourtzoglou, Joel D Schwartz, Marc G
Weisskopf, Steven J Melly, Yun Wang, Francesca Dominici, and
Antonella Zanobetti. Long-term pm2. 5 exposure and neuro-
logical hospital admissions in the northeastern united states.
Environmental health perspectives, 124(1):23–29, 2016.
[20]
N McCarthy. This is the global economic cost of air pollution.
In
World Economic Forum [online]. Cologny: World Economic
Forum, 2020.
[21]
Fiona Marshall, Mike Ashmore, Fiona Hinchcliffe, et al.
A hidden
threat to food production: Air pollution and agriculture in the
developing world
. International Institute for Environment and
Development., 1997.
[22]
Jennifer Burney and V Ramanathan. Recent climate and air
pollution impacts on indian agriculture.
Proceedings of the
National Academy of Sciences, 111(46):16319–16324, 2014.
[23]
Derek Elsom.
Smog alert: managing urban air quality
. Routledge,
2014.
[24]
Wei Ying Yi, Kin Ming Lo, Terrence Mak, Kwong Sak Leung, Yee
Leung, and Mei Ling Meng. A survey of wireless sensor network
based air pollution monitoring systems.
Sensors
, 15(12):31392–
31427, 2015.
[25]
Savina Singla, Divya Bansal, Archan Misra, and Gaurav Raheja.
Towards an integrated framework for air quality monitoring and
exposure estimation—a review.
Environmental monitoring and
assessment, 190(9):1–21, 2018.
[26]
Francesco Concas, Julien Mineraud, Eemil Lagerspetz, Samu
Varjonen, Xiaoli Liu, Kai Puolam
¨
aki, Petteri Nurmi, and Sasu
Tarkoma. Low-cost outdoor air quality monitoring and sensor
calibration: A survey and critical analysis.
ACM Transactions on
Sensor Networks (TOSN), 17(2):1–44, 2021.
[27]
Mannam Veera Narayana, Devendra Jalihal, and SM Nagendra.
Establishing a sustainable low-cost air quality monitoring setup:
A survey of the state-of-the-art. Sensors, 22(1):394, 2022.
[28]
Saverio De Vito, Elena Esposito, Ettore Massera, Fabrizio
Formisano, Grazia Fattoruso, Sergio Ferlito, Antonio Del Giudice,
Gerardo D’Elia, Maria Salvato, Tiziana Polichetti, et al. Crowd-
sensing iot architecture for pervasive air quality and exposome
monitoring: Design, development, calibration, and long-term
validation. Sensors, 21(15):5219, 2021.
[29]
Hocine Mokrani, Razika Lounas, Mohamed Tahar Bennai,
Dhai Eddine Salhi, and Rachid Djerbi. Air quality monitoring
using iot: A survey. In
2019 IEEE International Conference on
Smart Internet of Things (SmartIoT), pages 127–134. IEEE, 2019.
[30]
Balz Maag, Zimu Zhou, and Lothar Thiele. A survey on sensor
calibration in air pollution monitoring deployments.
IEEE Internet
of Things Journal, 5(6):4857–4870, 2018.
[31]
Jonathan E Thompson. Crowd-sourced air quality studies: A re-
view of the literature & portable sensors.
Trends in Environmental
Analytical Chemistry, 11:23–34, 2016.
[32]
Virendra Barot and Viral Kapadia. Air quality monitoring
systems using iot: A review. In
2020 International Conference
on Computational Performance Evaluation (ComPE)
, pages 226–
231. IEEE, 2020.
[33]
Saba Ameer, Munam Ali Shah, Abid Khan, Houbing Song, Carsten
Maple, Saif Ul Islam, and Muhammad Nabeel Asghar. Compar-
ative analysis of machine learning techniques for predicting air
quality in smart cities. IEEE Access, 7:128325–128338, 2019.
[34]
Movva Pavani and P Trinatha Rao. Urban air pollution mon-
itoring using wireless sensor networks: A comprehensive re-
view.
International Journal of Communication Networks and
Information Security, 9(3):439–449, 2017.
[35]
Akshara Kaginalkar, Shamita Kumar, Prashant Gargava, and Dev
Niyogi. Review of urban computing in air quality management
as smart city service: An integrated iot, ai, and cloud technology
perspective. Urban Climate, 39:100972, 2021.
[36]
Ntombikayise Koyana, Elisha Didam Markus, and Adnan M
Abu-Mahfouz. A survey of network and intelligent air pollution
monitoring in south africa. In
2019 International Multidisciplinary
Information Technology and Engineering Conference (IMITEC)
,
pages 1–7. IEEE, 2019.
[37]
R Kingsy Grace and S Manju. A comprehensive review of wireless
sensor networks based air pollution monitoring systems.
Wireless
Personal Communications, 108(4):2499–2515, 2019.
[38]
Vinit Lambey and AD Prasad. A review on air quality mea-
surement using an unmanned aerial vehicle.
Water, Air, & Soil
Pollution, 232(3):1–32, 2021.
[39]
Dharmendra Singh, Meenakshi Dahiya, Rahul Kumar, and Chin-
tan Nanda. Sensors and systems for air quality assessment
monitoring and management: A review.
Journal of Environmental
Management, 289:112510, 2021.
[40]
He Zhang and Ravi Srinivasan. A systematic review of air
quality sensors, guidelines, and measurement studies for indoor
air quality management. Sustainability, 12(21):9045, 2020.
[41]
Nurmadiha Osman, Mohd Faizal Jamlos, Fatimah Dzaharudin,
Aidil Redza Khan, You Kok Yeow, and Khairil Anuar Khairi.
Real-time and predictive analytics of air quality with iot system:
A review.
Recent Trends in Mechatronics Towards Industry 4.0
,
pages 107–116, 2022.
[42]
Raghavendra Khot and Vidya Chitre. Survey on air pol-
lution monitoring systems. In
2017 international conference
on innovations in information, embedded and communication
systems (ICIIECS), pages 1–4. IEEE, 2017.
[43]
Yuanfang He, Alexander Kuchansky, Sergiy Paliy, and Yevhenia
Shabala. Problems in air quality monitoring and assessment.
In
2021 IEEE International Conference on Smart Information
Systems and Technologies (SIST), pages 1–4. IEEE, 2021.
[44]
Min Huang and Xia Wu. A review of air quality monitoring
system based on crowdsensing. In
International Symposium
for Intelligent Transportation and Smart City
, pages 286–296.
Springer, 2019.
[45]
M Rogulski and A Badyda. Current trends in network based air
quality monitoring systems. In
IOP Conference Series: Earth and
Environmental Science
, volume 214, page 012085. IOP Publishing,
2019.
[46]
Ronnie Levin, Carolina L Zilli Vieira, Marieke H Rosenbaum,
Karyn Bischoff, Daniel C Mordarski, and Mary Jean Brown. The
urban lead (pb) burden in humans, animals and the natural
environment. Environmental research, 193:110377, 2021.
[47]
Carlyn J Matz, Marika Egyed, Robyn Hocking, Shayesta Seenun-
dun, Nick Charman, and Nigel Edmonds. Human health effects
of traffic-related air pollution (trap): a scoping review protocol.
Systematic reviews, 8(1):1–5, 2019.
[48]
Haneen Khreis, Mark Nieuwenhuijsen, Josias Zietsman, and Tara
Ramani. Traffic-Related Air Pollution. Elsevier, 2020.
[49]
Michael Greenstone and Claire Qing Fan. Introducing the air
quality life index. AQLI Annual Report, 2018.
[50]
Michael Greenstone and Qing Claire Fan. Aqli annual update.
2020.
[51]
Ju Zhang, Zhiyun Ouyang, Hong Miao, and Xiaoke Wang.
Ambient air quality trends and driving factor analysis in beijing,
1983–2007.
Journal of Environmental Sciences
, 23(12):2019–2028,
2011.
[52] Michael Greenstone and Qing Claire Fan. China fact sheet. 2020.
[53]
Michael Greenstone and Patrick Schwarz. Is china winning its
war on pollution.
Report from Energy Policy Institute at the
University of Chicago, 2018.
[54]
Michael Greenstone and Qing Claire Fan. North india fact sheet.
2019.
[55] Michael Greenstone and Qing Claire Fan. India fact sheet. 2020.
[56]
Michael Greenstone and Qing Claire Fan. India’s ‘war against
pollution’: An opportunity for longer lives. 2019.
[57]
Michael Greenstone and Qing Claire Fan. Southeast asia fact sheet.
2020.
[58]
Michael Greenstone and Qing Claire Fan. Indonesia fact sheet.
2020.
[59]
Michael Greenstone and Qing Claire Fan. Pakistan fact sheet.
2020.
[60]
Michael Greenstone and Qing Claire Fan. Bangladesh fact sheet.
2020.
[61]
Michael Greenstone and Qing Claire Fan. South korea fact sheet.
2019.
[62]
Michael Greenstone and Qing Claire Fan. Thailand fact sheet.
2019.
[63] Michael Greenstone and Claire Fan. Annual update.
[64]
Michael Greenstone and Qing Claire Fan. Europe fact sheet. 2020.
[65]
Per Erik Karlsson, Lars Robert Hole, Hans Tømmervik, and Elena
Kobets. Air pollution in the nordic countries, 2015.
[66]
Michael Greenstone and Qing Claire Fan. United states fact sheet.
2020.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

23
[67]
James Lents, Michael Walsh, K He, N Davis, M Osses, S Tolvett,
and L Hu. Handbook of air quality management.
Recuperado de:
www. aqbook. org, 2011.
[68]
Jos
´
e Usero, Antonio Lozano, Eva Vanderlinden, Juan Raez, Juan
Contreras, Benito Navarrete, and Hicham El Bakouri. Devel-
opment of the design of air quality monitoring networks and
its applications to no2 and o3 in seville, spain. In
Air Quality
.
IntechOpen, 2010.
[69]
David Gal
´
an Madruga. Importance of air quality networks in
controlling exposure to air pollution. In
Environmental Emissions
.
IntechOpen, 2020.
[70]
Abdullah Mofarrah and Tahir Husain. A holistic approach for
optimal design of air quality monitoring network expansion in an
urban area. Atmospheric Environment, 44(3):432–440, 2010.
[71]
Yu Kang, Jie Chen, Yang Cao, and Zhenyi Xu. A higher-order
graph convolutional network for location recommendation of an
air-quality-monitoring station. Remote Sensing, 13(8):1600, 2021.
[72]
Zhenyi Xu, Yu Kang, and Yang Cao. Emission stations lo-
cation selection based on conditional measurement gan data.
Neurocomputing, 388:170–180, 2020.
[73]
Zhengqiu Zhu, Bin Chen, Sihang Qiu, Rongxiao Wang, Liang Ma,
and Xiaogang Qiu. The gis-based approach for optimal design
of air quality monitoring network for management of chemical
clusters. In
Proceedings of the 3rd ACM SIGSPATIAL Workshop
on Emergency Management using, pages 1–8, 2017.
[74]
Hsun-Ping Hsieh, Shou-De Lin, and Yu Zheng. Inferring air
quality for station location recommendation based on urban big
data. In
Proceedings of the 21th ACM SIGKDD International
Conference on Knowledge Discovery and Data Mining
, pages
437–446, 2015.
[75]
Junyu Zheng, Xiaoqiong Feng, Panwei Liu, Liuju Zhong, and
Senchao Lai. Site location optimization of regional air quality
monitoring network in china: methodology and case study.
Journal of environmental monitoring, 13(11):3185–3195, 2011.
[76]
Dimosthenis A Sarigiannis and Michaela Saisana. Multi-objective
optimization of air quality monitoring.
Environmental monitoring
and assessment, 136(1):87–99, 2008.
[77]
Antonio Lozanoa, Jos
´
e Userob, Eva Vanderlindenb, Juan Raeza,
Juan Contrerasc, Benito Navarreteb, and Hicham El Bakourib.
Optimization of the design of air quality monitoring networks
and its application to no2 and o3 in seville, spain.
Air Quality
,
pages 49–64, 2010.
[78]
A Elkamel, E Fatehifar, M Taheri, MS Al-Rashidi, and A Lohi.
A heuristic optimization approach for air quality monitoring
network design with the simultaneous consideration of multiple
pollutants.
Journal of environmental management
, 88(3):507–516,
2008.
[79]
Yufang Hao and Shaodong Xie. Optimal redistribution of an urban
air quality monitoring network using atmospheric dispersion
model and genetic algorithm.
Atmospheric Environment
, 177:222–
233, 2018.
[80]
Tao Yu, Wen Wang, Pubu Ciren, and Rui Sun. An assessment of
air-quality monitoring station locations based on satellite observa-
tions.
International Journal of Remote Sensing
, 39(20):6463–6478,
2018.
[81]
Lidia Morawska, Phong K Thai, Xiaoting Liu, Akwasi Asumadu-
Sakyi, Godwin Ayoko, Alena Bartonova, Andrea Bedini, Fahe
Chai, Bryce Christensen, Matthew Dunbabin, et al. Applications
of low-cost sensing technologies for air quality monitoring and
exposure assessment: How far have they gone?
Environment
international, 116:286–299, 2018.
[82]
Alice Cavaliere, Federico Carotenuto, Filippo Di Gennaro, Beni-
amino Gioli, Giovanni Gualtieri, Francesca Martelli, Alessandro
Matese, Piero Toscano, Carolina Vagnoli, and Alessandro Zaldei.
Development of low-cost air quality stations for next generation
monitoring networks: Calibration and validation of pm2. 5 and
pm10 sensors. Sensors, 18(9):2843, 2018.
[83]
Matthias Budde, Matthias Berning, Mathias Busse, Takashi Miyaki,
and Michael Beigl. The teco envboard: A mobile sensor plat-
form for accurate urban sensing—and more. In
2012 Ninth
International Conference on Networked Sensing (INSS)
, pages
1–2. IEEE, 2012.
[84]
Piero Zappi, Elizabeth Bales, Jing Hong Park, William Griswold,
and T
ˇ
Simunic Rosing. The citisense air quality monitoring mobile
sensor node. In
Proceedings of the 11th ACM/IEEE conference on
information processing in sensor networks, Beijing, China
, pages
16–19. Citeseer, 2012.
[85]
Elizabeth Bales, Nima Nikzad, Nichole Quick, Celal Ziftci, Kevin
Patrick, and William Griswold. Citisense: Mobile air quality
sensing for individuals and communities design and deployment
of the citisense mobile air-quality system. In
2012 6th International
Conference on Pervasive Computing Technologies for Healthcare
(PervasiveHealth) and Workshops, pages 155–158. IEEE, 2012.
[86]
Nima Nikzad, Nakul Verma, Celal Ziftci, Elizabeth Bales, Nichole
Quick, Piero Zappi, Kevin Patrick, Sanjoy Dasgupta, Ingolf
Krueger, Tajana
ˇ
Simuni
´
c Rosing, et al. Citisense: Improving
geospatial environmental assessment of air quality using a
wireless personal exposure monitoring system. In
Proceedings of
the conference on Wireless Health, pages 1–8, 2012.
[87]
Vijay Sivaraman, James Carrapetta, Ke Hu, and Blanca Gallego
Luxan. Hazewatch: A participatory sensor system for monitoring
air pollution in sydney. In
38th Annual IEEE Conference on Local
Computer Networks-Workshops, pages 56–64. IEEE, 2013.
[88]
Ke Hu, Yan Wang, Ashfaqur Rahman, and Vijay Sivaraman.
Personalising pollution exposure estimates using wearable activity
sensors. In
2014 IEEE ninth international conference on intelligent
sensors, sensor networks and information processing (ISSNIP)
,
pages 1–6. IEEE, 2014.
[89]
Ke Hu, Ashfaqur Rahman, Hari Bhrugubanda, and Vijay Sivara-
man. Hazeest: Machine learning based metropolitan air pollution
estimation from fixed and mobile sensors.
IEEE Sensors Journal
,
17(11):3517–3525, 2017.
[90]
Matthias Budde, Matthias Berning, Mathias Busse, Takashi Miyaki,
and Michael Beigl. Handheld particulate matter measurements
with cots sensors.
Pervasive & ISWC 2012, Newcastle-upon-Tyne
,
2012.
[91]
Matthias Budde, Mathias Busse, and Michael Beigl. Investi-
gating the use of commodity dust sensors for the embedded
measurement of particulate matter. In
2012 Ninth International
Conference on Networked Sensing (INSS)
, pages 1–4. IEEE, 2012.
[92]
Balz Maag, Zimu Zhou, and Lothar Thiele. W-air: Enabling
personal air pollution monitoring on wearables.
Proceedings
of the ACM on Interactive, Mobile, Wearable and Ubiquitous
Technologies, 2(1):1–25, 2018.
[93]
Elizabeth Bales, Nima Nikzad, Nichole Quick, Celal Ziftci, Kevin
Patrick, and William G Griswold. Personal pollution monitoring:
mobile real-time air quality in daily life.
Personal and Ubiquitous
Computing, 23(2):309–328, 2019.
[94]
Ke Hu, Vijay Sivaraman, Blanca Gallego Luxan, and Ashfaqur
Rahman. Design and evaluation of a metropolitan air pollution
sensing system. IEEE Sensors Journal, 16(5):1448–1459, 2015.
[95]
Bart Elen, Jan Peters, Martine Van Poppel, Nico Bleux, Jan Theunis,
Matteo Reggente, and Arnout Standaert. The aeroflex: a bicycle
for mobile air quality measurements.
Sensors
, 13(1):221–240, 2013.
[96]
Srinivas Devarakonda, Parveen Sevusu, Hongzhang Liu, Ruilin
Liu, Liviu Iftode, and Badri Nath. Real-time air quality monitoring
through mobile sensing in metropolitan areas. In
Proceedings
of the 2nd ACM SIGKDD international workshop on urban
computing, pages 1–8, 2013.
[97]
Yi Gao, Wei Dong, Kai Guo, Xue Liu, Yuan Chen, Xiaojin
Liu, Jiajun Bu, and Chun Chen. Mosaic: A low-cost mobile
sensing system for urban air quality monitoring. In
IEEE
INFOCOM 2016-The 35th Annual IEEE International Conference
on Computer Communications, pages 1–9. IEEE, 2016.
[98]
Joshua S Apte, Kyle P Messier, Shahzad Gani, Michael Brauer,
Thomas W Kirchstetter, Melissa M Lunden, Julian D Marshall,
Christopher J Portier, Roel CH Vermeulen, and Steven P Hamburg.
High-resolution air pollution mapping with google street view
cars: exploiting big data.
Environmental science & technology
,
51(12):6999–7008, 2017.
[99]
Swati Dhingra, Rajasekhara Babu Madda, Amir H Gandomi,
Rizwan Patan, and Mahmoud Daneshmand. Internet of things
mobile–air pollution monitoring system (iot-mobair).
IEEE
Internet of Things Journal, 6(3):5577–5584, 2019.
[100]
Priyanka DeSouza, Amin Anjomshoaa, Fabio Duarte, Ralph
Kahn, Prashant Kumar, and Carlo Ratti. Air quality monitoring
using mobile low-cost sensors mounted on trash-trucks: Methods
development and lessons learned.
Sustainable Cities and Society
,
60:102239, 2020.
[101]
P Partheeban, H Prasad Raju, Ranganathan Rani Hemamalini,
and B Shanthini. Real-time vehicular air quality monitoring using
sensing technology for chennai. In
Transportation Research
, pages
19–28. Springer, 2020.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

24
[102]
Can Li, N Christina Hsu, and Si-Chee Tsay. A study on the
potential applications of satellite data in air quality monitoring
and forecasting.
Atmospheric Environment
, 45(22):3663–3675,
2011.
[103]
Meredith Fowlie, Edward Rubin, and Reed Walker. Bringing
satellite-based air quality estimates down to earth. In
AEA Papers
and Proceedings, volume 109, pages 283–88, 2019.
[104]
Jhoon Kim, Ukkyo Jeong, Myoung-Hwan Ahn, Jae H Kim,
Rokjin J Park, Hanlim Lee, Chul Han Song, Yong-Sang Choi,
Kwon-Ho Lee, Jung-Moon Yoo, et al. New era of air quality
monitoring from space: Geostationary environment monitoring
spectrometer (gems).
Bulletin of the American Meteorological
Society, 101(1):E1–E22, 2020.
[105]
Kerstin Stebel, Iwona S Stachlewska, Anca Nemuc, Jan Hor
´
alek,
Philipp Schneider, Nicolae Ajtai, Andrei Diamandi, Nina
Bene
ˇ
sov
´
a, Mihai Boldeanu, Camelia Botezan, et al. Samira-satellite
based monitoring initiative for regional air quality.
Remote
sensing, 13(11):2219, 2021.
[106] Daniel M Sullivan and Alan Krupnick. Using satellite data to fill
the gaps in the us air pollution monitoring network.
Resources
for the Future Working Paper, pages 18–21, 2018.
[107]
Matthew J Alvarado, Amy E McVey, Jennifer D Hegarty, Eben S
Cross, Christa A Hasenkopf, Richard Lynch, Edward J Kennelly,
Timothy B Onasch, Yewande Awe, Ernesto Sanchez-Triana, et al.
Evaluating the use of satellite observations to supplement ground-
level air quality data in selected cities in low-and middle-income
countries. Atmospheric Environment, 218:117016, 2019.
[108]
Jiayu Li, Huang Zhang, Chun-Ying Chao, Chih-Hsiang Chien,
Chang-Yu Wu, Cyuan Heng Luo, Ling-Jyh Chen, and Pratim
Biswas. Integrating low-cost air quality sensor networks with
fixed and satellite monitoring systems to study ground-level pm2.
5. Atmospheric Environment, 223:117293, 2020.
[109]
Heinz H Lettau. Physical and meteorological basis for math-
ematical models of urban diffusion processes. In
Proc. Symp.
on Multiple-Source Urban Diffusion Models, US Environ Protect.
Agency, pages 2–1, 1970.
[110]
Adel A Abdel-Rahman. On the atmospheric dispersion and
gaussian plume model. In
2nd International Conference on Waste
Management, Water Pollution, Air Pollution, Indoor Climate
(WWAI’08), pages 31–39, 2008.
[111]
AG Silveira, CS Mau
´
es, and CS Guimar
˜
aes. A rstudio-based
modeling program for dispersion of particulate matter emissions
using gaussian plume equation. In
Journal of Physics: Conference
Series, volume 1938, page 012013. IOP Publishing, 2021.
[112]
Bamidele Sunday Fakinle, Ebenezer Leke Odekanle, Oyinkan-
sola Serah Okunlola, Daniel Olawale Oke, Charity O Aremu,
Jacob Ademola Sonibare, and Funso Alaba Akeredolu. Spatial
dispersion modeling of air emissions from a farm using a gaussian
model. Environmental Quality Management, 2021.
[113]
Richard S Collett and Kehinde Oduyemi. Air quality modelling:
a technical review of mathematical approaches.
Meteorological
Applications, 4(3):235–246, 1997.
[114]
Dragana Popovic.
Air Quality: Models and Applications
. BoD–
Books on Demand, 2011.
[115]
Aaron Daly and Paolo Zannetti. Air pollution modeling–an
overview. Ambient air pollution, pages 15–28, 2007.
[116]
D Bruce Turner.
Workbook of atmospheric dispersion estimates:
an introduction to dispersion modeling. CRC press, 2020.
[117]
Ashok Kumar. Understanding the usepa’s aermod modeling
system.
Department of Civil Engineering, University of Toledo
,
2004.
[118]
Alan J Cimorelli, Steven G Perry, Akula Venkatram, Jeffrey C Weil,
Robert J Paine, Robert B Wilson, Russell F Lee, Warren D Peters,
and Roger W Brode. Aermod: A dispersion model for industrial
source applications. part i: General model formulation and
boundary layer characterization.
Journal of applied meteorology
,
44(5):682–693, 2005.
[119]
Steven G Perry, Alan J Cimorelli, Robert J Paine, Roger W Brode,
Jeffrey C Weil, Akula Venkatram, Robert B Wilson, Russell F Lee,
and Warren D Peters. Aermod: A dispersion model for industrial
source applications. part ii: Model performance against 17 field
study databases.
Journal of applied meteorology
, 44(5):694–708,
2005.
[120]
Duo-xing YANG, Mu-shui YANG, Xiao-hong ZHAO, Min LIU,
Ke-jia XING, and Lei QIU. Theory on aermod modeling system
[j]. Chemical industry and engineering, 2, 2005.
[121]
OAR US EPA. Air Quality Dispersion Modeling - Preferred and
Recommended Models, November 2016.
[122]
Steven G Perry. Ctdmplus: A dispersion model for sources near
complex topography. part i: Technical formulations.
Journal of
Applied Meteorology and Climatology, 31(7):633–645, 1992.
[123]
James O Paumier, Donna J Burns, and Steven G Perry. Ctdmplus:
A dispersion model for sources near complex topography. part ii:
Performance characteristics.
Journal of Applied Meteorology and
Climatology, 31(7):646–660, 1992.
[124]
Steven R Hanna, Lloyd L Schulman, Robert J Paine, Jonathan E
Pleim, and Mitchell Baer. Development and evaluation of the
offshore and coastal dispersion model.
Journal of the Air Pollution
Control Association, 35(10):1039–1047, 1985.
[125]
SR Hanna, LL Schulman, RJ Paine, and JE Pleim. Offshore and
coastal dispersion (ocd) model. users guide. Technical report,
Environmental Research and Technology, 1984.
[126]
Chunling Tang, Jason A Lynch, and Robin L Dennis. One multi-
media environmental system with linkage between meteorolo-
gy/hydrology/air quality models and water quality model. In
AGU Fall Meeting Abstracts, volume 2016, pages H33L–04, 2016.
[127]
Donna B Schwede and Gary G Lear. A novel hybrid approach
for estimating total deposition in the united states.
Atmospheric
Environment, 92:207–220, 2014.
[128]
DC Wong, J Pleim, R Mathur, F Binkowski, T Otte, R Gilliam,
G Pouliot, A Xiu, JO Young, and D Kang. Wrf-cmaq two-way
coupled system with aerosol feedback: software development and
preliminary results.
Geoscientific Model Development
, 5(2):299–
312, 2012.
[129]
SL Napelenok, DS Cohan, Mehmet T Odman, and S Tonse.
Extension and evaluation of sensitivity analysis capabilities in
a photochemical model.
Environmental Modelling & Software
,
23(8):994–999, 2008.
[130]
Daniel S Cohan and Sergey L Napelenok. Air quality response
modeling for decision support. Atmosphere, 2(3):407–425, 2011.
[131]
Sergey L Napelenok, Daniel S Cohan, Yongtao Hu, and Armis-
tead G Russell. Decoupled direct 3d sensitivity analysis for
particulate matter (ddm-3d/pm).
Atmospheric Environment
,
40(32):6112–6121, 2006.
[132]
Heather Simon, Luke C Valin, Kirk R Baker, Barron H Henderson,
James H Crawford, Sally E Pusede, James T Kelly, Kristen M Foley,
R Chris Owen, Ronald C Cohen, et al. Characterizing co and noy
sources and relative ambient ratios in the baltimore area using
ambient measurements and source attribution modeling.
Journal
of Geophysical Research: Atmospheres, 123(6):3304–3320, 2018.
[133]
R HF Kwok, Sergey L Napelenok, and Kirk R Baker. Implemen-
tation and evaluation of pm2. 5 source contribution analysis in
a photochemical model.
Atmospheric Environment
, 80:398–407,
2013.
[134]
Mark Estes, Jeremiah Johnson, Gary Wilson, Justin Bandoro, Kurt
Richman, Ling Huang, Ross Beardsley, and Greg Yarwood. Near-
real time exceptional event modeling final report. 2018.
[135]
Texas Commission on Environmental Quality. Hemispheric
comprehensive air quality model with extensions, enhancement
and testing. 2020.
[136] Texas Commission on Environmental Quality. Development and
testing of a hemispheric comprehensive air quality model with
extensions (camx) application. 2020.
[137]
Philip K Hopke et al.
Receptor modeling in environmental
chemistry. John Wiley & Sons, 1985.
[138]
Benjamin Loubet, Dominique Baisn
´
ee, Mathieu Cazaunau, Audrey
Fortems-Cheiney, Raluca Ciuraru, Cathy Clerbaux, Jean-Fran
c¸
ois
Doussin, Ga
¨
elle Dufour, Cristian Focsa, Christian George, et al.
Measuring air pollutant concentrations and fluxes. In
Agriculture
and Air Quality, pages 119–157. Springer, 2020.
[139]
Mohammad Hashem Askariyeh, Haneen Khreis, and Suriya
Vallamsundar. Air pollution monitoring and modeling. In
Traffic-Related Air Pollution, pages 111–135. Elsevier, 2020.
[140]
C Thomas Coulter. Epa-cmb8. 2 user’s manual (epa-452/r-04–011),
2004.
[141]
Philip K Hopke. A guide to positive matrix factorization. In
Workshop on UNMIX and PMF as Applied to PM2
, volume 5,
page 600, 2000.
[142]
Jacqueline Whalley and Sara Zandi. Particulate matter
sampling techniques and data modelling methods.
Air
Quality-Measurement and Modeling, pages 29–54, 2016.
[143]
MI Mead, OAM Popoola, GB Stewart, Peter Landshoff, M Calleja,
M Hayes, JJ Baldovi, MW McLeod, TF Hodgson, J Dicks, et al. The
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

25
use of electrochemical sensors for monitoring urban air quality
in low-cost, high-density networks.
Atmospheric Environment
,
70:186–203, 2013.
[144]
A. Clements and R. Vanderpool. Frms/fems and sensors: Com-
plementary approaches for determining ambient air quality.
[145]
Yu Zheng, Licia Capra, Ouri Wolfson, and Hai Yang. Urban
computing: Concepts, methodologies, and applications.
ACM
Transaction on Intelligent Systems and Technology
, October 2014.
[146]
Yu Zheng, Furui Liu, and Hsun-Ping Hsieh. U-air: When urban
air quality inference meets big data. In
Proceedings of the 19th
ACM SIGKDD international conference on Knowledge discovery
and data mining, pages 1436–1444, 2013.
[147]
Yu Zheng, Xuxu Chen, Qiwei Jin, Yubiao Chen, Xiangyun Qu,
Xin Liu, Eric Chang, Wei-Ying Ma, Yong Rui, and Weiwei Sun.
A cloud-based knowledge discovery system for monitoring fine-
grained air quality. MSR-TR-2014–40, Tech. Rep., 2014.
[148]
Yu Zheng, Xiuwen Yi, Ming Li, Ruiyuan Li, Zhangqing Shan,
Eric Chang, and Tianrui Li. Forecasting fine-grained air qual-
ity based on big data. In
Proceedings of the 21th ACM
SIGKDD international conference on knowledge discovery and
data mining, pages 2267–2276, 2015.
[149]
Chao Zhang, Yu Zheng, Xiuli Ma, and Jiawei Han. As-
sembler: Efficient discovery of spatial co-evolving patterns in
massive geo-sensory data. In
Proceedings of the 21th ACM
SIGKDD international conference on Knowledge discovery and
data mining, pages 1415–1424, 2015.
[150]
Hoang Nguyen, Wei Liu, and Fang Chen. Discovering conges-
tion propagation patterns in spatio-temporal traffic data.
IEEE
Transactions on Big Data, 3(2):169–180, 2016.
[151]
Leland Gerson Neuberg. Causality: models, reasoning, and
inference, by judea pearl, cambridge university press, 2000.
Econometric Theory, 19(4):675–685, 2003.
[152]
Julie Yixuan Zhu, Yu Zheng, Xiuwen Yi, and Victor OK Li. A
gaussian bayesian model to identify spatio-temporal causalities for
air pollution based on urban big data. In
2016 IEEE Conference on
Computer Communications Workshops (INFOCOM WKSHPS)
,
pages 3–8. IEEE, 2016.
[153]
Jiawei Han, Jian Pei, Behzad Mortazavi-Asl, Helen Pinto, Qiming
Chen, Umeshwar Dayal, and Meichun Hsu. Prefixspan: Mining
sequential patterns efficiently by prefix-projected pattern growth.
In
proceedings of the 17th international conference on data
engineering, pages 215–224. Citeseer, 2001.
[154]
Lionel Barnett, Adam B Barrett, and Anil K Seth. Granger causal-
ity and transfer entropy are equivalent for gaussian variables.
Physical review letters, 103(23):238701, 2009.
[155]
Katerina Hlav
´
a
ˇ
ckov
´
a-Schindler, Milan Palu
ˇ
s, Martin Vejmelka,
and Joydeep Bhattacharya. Causality detection based on
information-theoretic approaches in time series analysis.
Physics
Reports, 441(1):1–46, 2007.
[156]
Melanie S Hammer, Aaron van Donkelaar, Chi Li, Alexei Lya-
pustin, Andrew M Sayer, N Christina Hsu, Robert C Levy,
Michael J Garay, Olga V Kalashnikova, Ralph A Kahn, et al.
Global estimates and long-term trends of fine particulate matter
concentrations (1998–2018).
Environmental Science & Technology
,
54(13):7879–7890, 2020.
[157]
Avraham Ebenstein, Maoyong Fan, Michael Greenstone, Guojun
He, and Maigeng Zhou. New evidence on the impact of sustained
exposure to air pollution on life expectancy from china’s huai
river policy.
Proceedings of the National Academy of Sciences
,
114(39):10384–10389, 2017.
[158]
Susanne E Bauer, Ulas Im, Keren Mezuman, and Chloe Y Gao.
Desert dust, industrialization, and agricultural fires: Health
impacts of outdoor air pollution in africa.
Journal of Geophysical
Research: Atmospheres, 124(7):4104–4120, 2019.
[159]
Michael R Giordano et al. Afriqair ’s mission towards cleaner air
for africa and a call to action. Clean Air Journal, 31(1):1–3, 2021.
[160]
Stephen T Odonkor and Tahiru Mahami. Knowledge, attitudes,
and perceptions of air pollution in accra, ghana: A critical survey.
Journal of environmental and public health, 2020, 2020.
[161]
A Kofi Amegah. Proliferation of low-cost sensors. what prospects
for air pollution epidemiologic research in sub-saharan africa?
Environmental pollution, 241:1132–1137, 2018.
[162]
Abosede S Alli, Sierra N Clark, Allison Hughes, James Nimo,
Josephine Bedford-Moses, Solomon Baah, Jiayuan Wang, Jose
Vallarino, Ernest Agyemang, Benjamin Barratt, et al. Spatial-
temporal patterns of ambient fine particulate matter (pm2. 5) and
black carbon (bc) pollution in accra.
Environmental Research
Letters, 16(7):074013, 2021.
[163] R Subramanian and Rebecca M Garland. The powerful potential
of low-cost sensors for air quality research in africa.
Clean Air
Journal, 31(1):1–1, 2021.
[164]
A Kofi Amegah. Leveraging low-cost air quality sensors and
machine learning techniques for air pollution assessment and
prediction in urban ghana. In
ISEE Conference Abstracts
, volume
2021, 2021.
[165]
Christian Sewor and Akua A Obeng. The ghana urban air quality
project (ghair): Bridging air pollution data gaps in ghana.
Clean
Air Journal, 31(1):1–2, 2021.
[166]
James Carrapetta. Haze watch: Design of a wireless sensor board
for measuring air pollution.
Unversity of New South Wales,
Sydney, Australia, Tech. Rep, 2010.
[167]
J Liu. Calibration of a wireless sensor board for measuring air
pollution.
Unversity of New South Wales, Sydney, Australia, Tech.
Rep, 2012.
[168]
M Kelly. Design of a wireless sensor board for an air pollution
monitoring system.
Unversity of New South Wales, Sydney,
Australia, Tech. Rep, 2012.
[169]
Zhang Hailian. Haze watch project: Communication between
sensors and mobile phone.
[170]
Jiang Junjie. Communication between sensors and mobile phone.
2011.
[171]
Nikolaus Youdale. Haze watch: Database server and mobile
applications for measuring and evaluating air pollution exposure.
Unversity of New South Wales, Sydney, Australia, Tech. Rep
,
2010.
[172]
D Lu.
Android interface for pollution monitoring system
. PhD
thesis, Thesis B Report. School of Electrical and Telecommunica-
tions Engineering . . . , 2011.
[173]
Amanda Chow. Haze watch—data modelling and visualisation
using google maps.
Unversity of New South Wales, Sydney,
Australia, Tech. Rep, 2010.
[174]
Audrey De Nazelle, Edmund Seto, David Donaire-Gonzalez,
Michelle Mendez, Jaume Matamala, Mark J Nieuwenhuijsen,
and Michael Jerrett. Improving estimates of air pollution ex-
posure through ubiquitous sensing technologies.
Environmental
pollution, 176:92–99, 2013.
[175]
Nuria Castell, Franck R Dauge, Philipp Schneider, Matthias Vogt,
Uri Lerner, Barak Fishbain, David Broday, and Alena Bartonova.
Can commercial low-cost sensor platforms contribute to air quality
monitoring and exposure estimates?
Environment international
,
99:293–302, 2017.
[176]
Milena Jova
ˇ
sevi
´
c-Stojanovi
´
c, Alena Bartonova, Du
ˇ
san Topalovi
´
c,
Ivan Lazovi
´
c, Boris Pokri
´
c, and Zoran Ristovski. On the use
of small and cheaper sensors and devices for indicative citizen-
based monitoring of respirable particulate matter.
Environmental
Pollution, 206:696–704, 2015.
[177]
Nuria Castell, Philipp Schneider, Sonja Grossberndt, Mirjam F
Fredriksen, Gabriela Sousa-Santos, Mathias Vogt, and Alena
Bartonova. Localized real-time information on outdoor air quality
at kindergartens in oslo, norway using low-cost sensor nodes.
Environmental research, 165:410–419, 2018.
[178]
Barak Fishbain, Uri Lerner, Nuria Castell, Tom Cole-Hunter,
Olalekan Popoola, David M Broday, Tania Martinez I
˜
niguez, Mark
Nieuwenhuijsen, Milena Jovasevic-Stojanovic, Dusan Topalovic,
et al. An evaluation tool kit of air quality micro-sensing units.
Science of the Total Environment, 575:639–648, 2017.
[179]
Radim J Sram, Miroslav Dostal, Helena Libalova, Pavel Rossner,
Andrea Rossnerova, Vlasta Svecova, Jan Topinka, and Alena
Bartonova. The european hot spot of b [a] p and pm2. 5
exposure—the ostrava region, czech republic: health research
results. International Scholarly Research Notices, 2013, 2013.
[180]
Nuria Castell and Sonja Grossberndt. Oslo citizens’ observatory.
results from the oslo empowerment initiative as part of the citi-
sense project. NILU rapport, 2018.
[181]
Philipp Schneider, Nuria Castell, Matthias Vogt, Franck R Dauge,
William A Lahoz, and Alena Bartonova. Mapping urban air
quality in near real-time using observations from low-cost sensors
and model information.
Environment international
, 106:234–247,
2017.
[182]
William A Lahoz and Philipp Schneider. Data assimilation:
making sense of earth observation.
Frontiers in Environmental
Science, 2:16, 2014.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

26
[183]
Shai Moshenberg, Uri Lerner, and Barak Fishbain. Spectral meth-
ods for imputation of missing air quality data.
Environmental
Systems Research, 4(1):1–13, 2015.
[184]
Radim J Sram et al. Molecular epidemiology and air pollution.
Chapters, 2015.
[185]
Barak Fishbain and Erick Moreno-Centeno. Self calibrated wireless
distributed environmental sensory networks.
Scientific reports
,
6(1):1–10, 2016.
[186]
Fadi Kizel, Yael Etzion, Rakefet Shafran-Nathan, Ilan Levy, Barak
Fishbain, Alena Bartonova, and David M Broday. Node-to-
node field calibration of wireless distributed air pollution sensor
network. Environmental Pollution, 233:900–909, 2018.
[187]
Saverio De Vito, Elena Esposito, Nuria Castell, Philipp Schneider,
and Alena Bartonova. On the robustness of field calibration for
smart air quality monitors.
Sensors and Actuators B: Chemical
,
310:127869, 2020.
[188]
David M Holstius, A Pillarisetti, KR Smith, and EJAMT Seto.
Field calibrations of a low-cost aerosol sensor at a regulatory mon-
itoring site in california.
Atmospheric Measurement Techniques
,
7(4):1121–1131, 2014.
[189]
Du
ˇ
san B Topalovi
´
c, Milo
ˇ
s D Davidovi
´
c, Maja Jovanovi
´
c, Alena
Bartonova, Zoran Ristovski, and Milena Jova
ˇ
sevi
´
c-Stojanovi
´
c. In
search of an optimal in-field calibration method of low-cost gas
sensors for ambient air pollutants: Comparison of linear, mul-
tilinear and artificial neural network approaches.
Atmospheric
Environment, 213:640–658, 2019.
[190]
Maja V Jovanovic, Jasmina Z Savic, Farhad Salimi, Svetlana
Stevanovic, Reece A Brown, Milena Jovasevic-Stojanovic, Dragan
Manojlovic, Alena Bartonova, Steven Bottle, and Zoran D Ris-
tovski. Measurements of oxidative potential of particulate matter
at belgrade tunnel; comparison of bpeanit, dtt and dcfh assays.
International journal of environmental research and public health
,
16(24):4906, 2019.
[191]
Ilan Levy. A national day with near zero emissions and its effect
on primary and secondary pollutants.
Atmospheric Environment
,
77:202–212, 2013.
[192]
David M Broday, Citi-Sense Project Collaborators, et al. Wireless
distributed environmental sensor networks for air pollution
measurement—the promise and the current reality.
Sensors
,
17(10):2263, 2017.
[193]
Johanna Amalia Robinson, David Kocman, Milena Horvat, and
Alena Bartonova. End-user feedback on a low-cost portable air
quality sensor system—are we there yet?
Sensors
, 18(11):3768,
2018.
[194]
Sonja Grossberndt, Philipp Schneider, Hai-Ying Liu, Mirjam
Fredriksen, Nuria Castell, Panagiota Syropoulou, and Alena
Bartonova. Public perception of urban air quality using vol-
unteered geographicinformation services. 2020.
[195]
Boris Pokric, Srdan Krco, Dejan Drajic, Maja Pokric, Ivan Jo-
kic, and Milena Jovaevic Stojanovic. ekonet-environmental
monitoring using low-cost sensors for detecting gases, partic-
ulate matter, and meteorological parameters. In
2014 Eighth
International Conference on Innovative Mobile and Internet
Services in Ubiquitous Computing, pages 421–426. IEEE, 2014.
[196]
Hai-Ying Liu, Mike Kobernus, David Broday, and Alena
Bartonova. A conceptual approach to a citizens’ observatory–
supporting community-based environmental governance.
Environmental Health, 13(1):1–13, 2014.
[197]
David Hasenfratz, Olga Saukh, Christoph Walser, Christoph
Hueglin, Martin Fierz, and Lothar Thiele. Pushing the spatio-
temporal resolution limit of urban air pollution maps. In
2014 IEEE International Conference on Pervasive Computing and
Communications (PerCom), pages 69–77. IEEE, 2014.
[198]
David Hasenfratz, Olga Saukh, Christoph Walser, Christoph
Hueglin, Martin Fierz, and Lothar Thiele. Revealing the limits of
spatio-temporal high-resolution pollution maps. In
Proceedings
of the 11th ACM Conference on Embedded Networked Sensor
Systems, pages 1–2, 2013.
[199]
Adrian Arfire, Ali Marjovi, and Alcherio Martinoli. Model-based
rendezvous calibration of mobile sensor networks for monitoring
air quality. In 2015 IEEE SENSORS, pages 1–4. IEEE, 2015.
[200]
Ali Marjovi, Adrian Arfire, and Alcherio Martinoli. High reso-
lution air pollution maps in urban environments using mobile
sensor networks. In
2015 International Conference on Distributed
Computing in Sensor Systems, pages 11–20. IEEE, 2015.
[201]
Dai-Hua Tsai, Idris Guessous, Michael Riediker, Fred Paccaud,
Jean-Michel Gaspoz, Jean-Marc Theler, Gerard Waeber, Peter Vol-
lenweider, and Murielle Bochud. Short-term effects of particulate
matters on pulse pressure in two general population studies.
Journal of hypertension, 33(6):1144–1152, 2015.
[202]
Federico Karagulian, Claudio A Belis, Carlos Francisco C Dora,
Annette M Pr
¨
uss-Ust
¨
un, Sophie Bonjour, Heather Adair-Rohani,
and Markus Amann. Contributions to cities’ ambient particulate
matter (pm): A systematic review of local source contributions at
global level. Atmospheric environment, 120:475–483, 2015.
[203]
Lei Jiang and Ling Bai. Spatio-temporal characteristics of urban
air pollutions and their causal relationships: Evidence from beijing
and its neighboring cities. Scientific reports, 8(1):1–12, 2018.
[204]
Clive WJ Granger. Investigating causal relations by econometric
models and cross-spectral methods.
Econometrica: journal of the
Econometric Society, pages 424–438, 1969.
[205]
Hong Wang, Jiayu Xu, Meng Zhang, Yuanqin Yang, Xiaojing
Shen, Yaqiang Wang, Dong Chen, and Jianping Guo. A study
of the meteorological causes of a prolonged and severe haze
episode in january 2013 over central-eastern china.
Atmospheric
Environment, 98:146–157, 2014.
[206]
KB Gurumoorthy, SP Vimal, N Sathish Kumar, and M Kasisel-
vanathan. Air pollution hotspot detection and identification of
their source trajectory. In
Journal of Physics: Conference Series
,
volume 1917, page 012029. IOP Publishing, 2021.
[207]
Yousif Alyousifi, Kamarulzaman Ibrahim, Wei Kang, and Wan
Zawiah Wan Zin. Robust empirical bayes approach for markov
chain modeling of air pollution index.
Journal of Environmental
Health Science and Engineering, 19(1):343–356, 2021.
[208]
S Meivel, M Mahesh, S Mohnish, and S Mukesh Pandian. Air
pollution monitoring system using iot and artificial intelligence.
International Journal of Modern Agriculture
, 10(2):3085–3089,
2021.
[209]
P Ajitha, Nikitha Reddy, Swetha Srinivasan, A Sivasangari,
RM Gomathi, and E Brumancia. Prediction of air quality based
on supervised learning. In
2021 5th International Conference on
Trends in Electronics and Informatics (ICOEI)
, pages 1261–1267.
IEEE, 2021.
[210]
Yulia Yamineva and Seita Romppanen. Is law failing to address
air pollution? reflections on international and eu developments.
Review of European, comparative & international environmental
law, 26(3):189–200, 2017.
[211]
Alfonso Garcia-de Prado, Guadalupe Ortiz, Juan Boubeta-Puig,
and David Corral-Plaza. Air4people: a smart air quality monitor-
ing and context-aware notification system.
Journal of Universal
Computer Science, 24(7):846–863, 2018.
[212] Sofoklis Kyriazakos, Mihail Mihaylov, Bayu Anggorojati, Albena
Mihovska, Razvan Craciunescu, Octavian Fratu, and Ramjee
Prasad. ewall: an intelligent caring home environment offering
personalized context-aware applications based on advanced
sensing.
Wireless Personal Communications
, 87(3):1093–1111,
2016.
[213]
Divya Lohani and Debopam Acharya. Smartvent: A context
aware iot system to measure indoor air quality and ventilation
rate. In
2016 17th IEEE International Conference on Mobile Data
Management (MDM), volume 2, pages 64–69. IEEE, 2016.
[214]
Meruyert Nurgazy, Arkady Zaslavsky, Prem Prakash Jayaraman,
Sylvain Kubler, Karan Mitra, and Saguna Saguna. Cavisap:
Context-aware visualization of outdoor air pollution with iot
platforms. In
2019 International Conference on High Performance
Computing & Simulation (HPCS), pages 84–91. IEEE, 2019.
[215]
Fran Casino, Constantinos Patsakis, Edgar Batista, Frederic Borr
`
as,
and Antoni Mart
´
ınez-Ballest
´
e. Healthy routes in the smart city: A
context-aware mobile recommender.
IEEE Software
, 34(6):42–47,
2017.
[216]
Daniel Sch
¨
urholz, Meruyert Nurgazy, Arkady Zaslavsky,
Prem Prakash Jayaraman, Sylvain Kubler, Karan Mitra, and
Saguna Saguna. Myaqi: Context-aware outdoor air pollution
monitoring system. In
Proceedings of the 9th International
Conference on the Internet of Things, pages 1–8, 2019.
[217]
Tamara Belyakhina, Arkady Zaslavsky, Karan Mitra, Saguna Sa-
guna, and Prem Prakash Jayaraman. Discpaq: Distributed context
acquisition and reasoning for personalized indoor air quality
monitoring in iot-based systems. In
Internet of Things, Smart
Spaces, and Next Generation Networks and Systems
, pages 75–
86. Springer, 2017.
[218]
Kwok Wai Tham. Indoor air quality and its effects on humans—a
review of challenges and developments in the last 30 years.
Energy
and Buildings, 130:637–650, 2016.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1

27
[219]
MY Zakaria Abouleish. Indoor air quality and covid-19.
Public
Health, 191:1, 2021.
[220]
Samuel Dom
´
ınguez-Amarillo, Jesica Fern
´
andez-Ag
¨
uera, Sonia
Cesteros-Garc
´
ıa, and Roberto Alonso Gonz
´
alez-Lezcano. Bad
air can also kill: residential indoor air quality and pollutant
exposure risk during the covid-19 crisis.
International Journal
of Environmental Research and Public Health, 17(19):7183, 2020.
[221]
Azra Kenarkoohi, Zahra Noorimotlagh, Shahab Falahi, Ali Amar-
loei, Seyyed Abbas Mirzaee, Iraj Pakzad, and Elham Bastani.
Hospital indoor air quality monitoring for the detection of
sars-cov-2 (covid-19) virus.
Science of The Total Environment
,
748:141324, 2020.
[222]
Jagriti Saini, Maitreyee Dutta, and Gon
c¸
alo Marques. Indoor air
quality monitoring systems and covid-19.
Emerging Technologies
During the Era of COVID-19 Pandemic, 348:133, 2021.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 26 April 2022 doi:10.20944/preprints202204.0232.v1