[ Research Article ]
Asian Journal of Atmospheric Environment - Vol. 15, No. 4, pp.78-92
ISSN: 1976-6912 (Print) 2287-1160 (Online)
Print publication date 31 Dec 2021
Received 12 Aug 2021 Revised 08 Oct 2021 Accepted 08 Dec 2021

# Measurement of Traffic-Related Air Pollution in Peshawar, Pakistan : A Pilot Study

Mohsin Khan* ; Mohammad Abdul Aziz Irfan ; Najeeb Ullah1)
Department of Mechanical Engineering, University of Engineering & Technology (UET) Peshawar, Peshawar 25120, Khyber Pakhtunkhwa, Pakistan
1)U.S. Pakistan Center for Advanced Studies in Energy (USPCAS-E), UET Peshawar, Phase 5 Hayatabad, Khyber Pakhtunkhwa, Pakistan

Correspondence to: * Tel: +92 306 0004025 E-mail: Mohsinsafi67@gmail.com

Copyright © 2021 by Asian Association for Atmospheric Environment
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

## Abstract

This pilot study measured Traffic-Related Air Pollution (TRAP) and calculated the corresponding Air Quality Index (AQI) in Peshawar. Using Libelium wireless sensors, the research measured outdoor TRAP and monitored indoor air quality for 48 days. The maximum outdoors daily mean concentration was 47 μg m-3 for PM1, 90 μg m-3 for PM2.5, 356 μg m-3 for PM10, 258 ppb for SO2, and 219 ppb for NO2, respectively. This corresponds to PM2.5 AQI of 158 (Unhealthy), PM10 AQI of 148 (Unhealthy for Sensitive Groups; USG), SO2 AQI of 181 (Unhealthy), and NO2 AQI of 123 (USG). The maximum daily average concentration for the indoor condition was 31 μg m-3 for PM1, 49 μg m-3 for PM2.5, 78 μg m-3 for PM10, 465 ppb for SO2, and 247 ppb for NO2, respectively. The corresponding AQI was 135 (USG) for PM2.5, 62 (Moderate) for PM10, 254 (Very Unhealthy) for SO2, and 129 (USG) for NO2. Data analysis shows that about 73% of the overall indoor AQI falls in the category of “USG”, while SO2 was the largest contributor to overall AQI. The study concludes that indoor AQI was slightly better than outdoor AQI because of the distance and height from the outdoor location. Moreover, Pakistan’s AQI for PM2.5 exceeds WHO’s 24-hours limit; however, it was relatively better by 23%, 65%, and 170% compared to China, India, and Bangladesh, respectively. In contrast, AQI for SO2 and NO2 was poor as compared to the same countries. The concentration and AQI for traffic-related air pollutants remain unhealthy and sometimes becomes hazardous, which means the sensitive groups are at greater risk.

## Keywords:

Air Quality Index (AQI), Climate change, Environmental protection, Particulate matter, Traffic-Related Air Pollution (TRAP)

## 1. INTRODUCTION

Urban air pollution is of growing concern, a larger part of which is attributed to vehicular traffic emissions. In Pakistan, within two decades, from 1991 to 2012, the number of automobile vehicles has increased roughly by 530% (Sánchez-Triana et al., 2014) from around 2 million to 10.6 million. Rapid growth in demand for motor vehicles and their excessive use has led to elevated traffic emissions, which generally pose an adverse health risk to the residents in the city. The energy used by the transportation section in Pakistan accounts for 31.4% of the total energy consumption (Khan and Yasmin, 2014). Similarly, oil in the transport sector accounts for 77% (Ministry of Finance, 2019a). In 2008, 85% of PM2.5 was due to road transport from fossil fuel combustion, while 12.2% was due to combustion in the manufacturing industry, and the rest 2.3% included other combustion sources. Similarly, 72% of PM10 arose due to fossil fuel combustion in road transport, 21.6% due to combustion in the manufacturing industry, and 6.4% included other combustion sources (Sánchez-Triana et al., 2014).

Major traffic-related air pollution causes are fossil fuel combustion, poor vehicle maintenance, and low-quality fuel consumption. A decade-long study conducted by Pérez-Martínez et al., (2015) to assess traffic emissions of Sao Palo, Brazil, concluded that 63% of CO2 is attributed to the consumption of gasoline and ethanol while 83% of NOx is due to diesel consumption. It was observed from the two-years study conducted by Linares and Díaz (2010) in Madrid that PM2.5 is the most significant traffic air pollutant among PM10, NOx, SO2, NO2, and O3. Other pollutants originate from light-duty vehicles (LDVs), which use gasoline and ethanol as fuel and are the cause of CO and volatile organic compounds (VOC); also, heavy-duty vehicles (HDVs) generate NOx and particulate matter (PM) (Pérez-Martínez et al., 2015).

Dispersion of PM1 and PM2.5 was investigated by Wu et al., (2002) and found that concentration reduces with an increase in the height of 2-79 m, while a difference up to 10% in the horizontal direction for 280 m was noticed. In contrast, the maximum concentration exists between 2 to 3-meter height from the Earth’s surface. It was elucidated by Kim et al., (2004) that TRAP has a high concentration for schools within a 300-meter circle compared to schools outside of this range.

Traffic-related air pollution has health, economic and environmental impact. Newman et al., (2014) claimed that 18.6% of children having asthma were readmitted to the hospital because of exposure to TRAP. Ten schools in high traffic areas near busy roads at San Francisco Bay in 2011 were investigated by Kim et al., (2004), and it was found that asthma has a linear relation with TRAP in children. Linares and Díaz (2010) concluded that hospital readmission for age 75 or more was increased due to TRAP. Also, an increase in blood pressure, variation in heart rate, myocardial infarction, stroke, ischemic disease, coronary atherosclerosis, arrhythmias, and infarction are greatly related to fine particulate matter PM2.5 exposure. The study by Costa et al., (2014) suggests that TRAP is a key contributor to air pollution and damaging effect on behaviour abnormalities, neuroinflammation, and oxidative stress. Further, high emission concentration contributes to developmental disabilities, hyper activities, autism, neurotoxicity, and decreased cognitive functions. The consequences of particulate matter on lung function, respiratory problems, and pulse rate have been investigated in Delhi by Kesavachandran et al., (2015), which shows that increase in the concentration of PM1 and PM2.5 has declined the forced respiratory volume (1-sec FEV) and peak respiratory flow rate (PEFR). To know the consequences of lead (Pb) from traffic air pollution on the nearest roadside soil and plants, Ahmad et al., (2019a) have tested roadside soil and leaf samples from five different locations. The test reveals that lead (Pb) was present from 28.8 to 44.8 mg kg-1 in concentration for soil while the concentration of 3.48 to 5.93 mg kg-1 in the leaves. By studying the effect of reducing TRAP in Rome by Cesaroni et al., (2012), from 2001 to 2005, it was revealed that 3.4 days of life was gained per person for citizens living near high traffic zones. Pascal et al., (2013) investigated filthy air’s economic and health consequences in twenty-five European cities by estimating air and health figures. After a controlled decrement of 2.5 μg m-3 in the air, it was observed from health data that a massive € 14 billion could be saved annually. Lin et al., (2018) found that the cost of energy increases in providing cleaner air to buildings, as the ambient air has been contaminated by particle matter due to high economic activities assisted by transportation. Investigation of the effects of traffic policies implementation in 2018 in Beijing, China by Fontes et al., (2018) reveals that the first traffic curtailment reduced PM2.5 by 5% from 95.4 μg m-3 to 91.1 μg m-3 while the second traffic policy implementation has increased concentration by 16%, from 102.4 μg m-3 to 121.6 μg m-3. The measure restricted certain vehicles in a specific area for a particular time. Anwar et al., (2021) investigated air pollution of various sources, including vehicular traffic in China, India, and Pakistan. The study compiles that air pollution mitigation is a multidimensional approach, including making legislation, implementing laws and rules, establishing policies, implementing new technologies, upgrading emissions standards, and continuously monitoring air pollution. Air pollution, especially particulate matter from various sources, including traffic, reduces visibility claimed by Zhao et al., (2013) and causes regional haze US-EPA (2006), affecting tourism and air traffic. Huang et al., (2016), in their 12 years investigation, concluded that visibility had been greatly deteriorated because of worsening air quality. Due to the absorption of sunlight by particulate matter and the excess of greenhouse gases, climate changes are prominent, and an increase in global average temperatures is obvious, Lelieveld et al., (2019) concluded. The rise in temperature poses a greater threat to food security as the investigation by Asseng et al., (2014) reveals that there is a reduction in wheat yield by 6% with an increase of 1°C.

Numerous methods have been implemented to study TRAP. Berkowicz et al., (2006) used the Danish Operational Street Pollution Model (OSPM) of estimation for traffic emission (NOx and CO) at the Copenhagen University building based on the COPERT model. Other studies, such as Zanella et al., (2014), have conducted a pilot study at Padova, Italy, about new IoT technology, for the street lighting system and measurement of environmental parameters. Mobile Crowd Sensing (MCS) was another IoT technique coined by Montori et al., (2018), which collected corporate, state, and end-user data for analysis. Another research by Ahmad et al., (2019b) determined the particulate matter PM10 and Pb concentration at Upper Dir and Charsadda using Reference Ambient Air Sampler and graphite furnace atomic spectroscopy. Similarly, an investigation discovered the temperature, humidity, and CO2 level using IoT, where data from the transmitter node has been sent to the receiver node and stored in a customized excel sheet (Shah and Mishra, 2016). To measure the concentration of CO2, NOx, and PM10, non-dispersive infrared photometry, chemiluminescence, and automatic sampling of beta radiation methods were used respectively by Pérez-Martínez et al., (2015). The correlation of exposure to TRAP and hospital readmission was analyzed by Newman et al., (2014) for a period of 12 months employing logistic regression and the Cox proportional hazard regression model (a regression model for evaluating simultaneously the influence of multiple factors on patient survival or the rate of a particular event happening (e.g., infection, death)) and found that TRAP has a positive effect on readmission to hospital.

Not many studies assessed TRAP, as most of them worked on the interconnection of health with air pollution (Pascal et al., 2013) and focused on the impact of TRAP on asthma (Newman et al., 2014) and other health issues (Ahmad et al., 2019a; Kesavachandran et al., 2015; Costa et al., 2014; Cesaroni et al., 2012; Linares and Díaz, 2010; Kim et al., 2004). There is hardly any research for monitoring TRAP based on contemporary IoT systems. However, traffic-related air pollution is the primary concern in this research. It is being analyzed by measuring and monitoring the concentration of parameters responsible for the degradation of air quality. Mitigating traffic-related air pollution (TRAP) is a formidable challenge, and consistent long-term commitment to environmental protection is indispensable. This pilot study covers the aspects of measurement, monitoring, processing, and reporting of data related to TRAP in the city hotspot, in particular at BRT Bus Station at Peshawar University, opposite to Khyber Teaching Hospital (KTH) Peshawar for outdoor investigation while at the University of Engineering and Technology Peshawar for indoor infiltration. Wireless sensors technology was deployed on the road for data acquisition for the concentration of traffic air pollution by measuring the concentration of PM1, PM2.5, PM10, NO2, SO2, CO2, and meteorological parameters (pressure, temperature, and humidity). The measured concentration was transmitted to Amazon Web Services (AWS) for storage, analysis, monitoring, and reporting purposes. The Air Quality Index (AQI) was calculated from the pollutant’s concentration. The study of TRAP remains a nascent topic in Pakistan, and very little data is available in this regard. The current study is one of the first efforts to measure TRAP in Peshawar, Pakistan. This data analysis would provide a source to make data-based decisions, an intelligent, reliable, and acceptable ways to build strategies for limiting TRAP in the city.

## 2. METHODOLOGY

The current research consists of these main parts (a) Setting up Libelium’s Smart Cities Pro Model along with sensor probes (b) Sensors deployment at the site (c) Using Libelium Cloud Bridge as buffer storage (d) Using AWS IoT and AWS DynamoDB for sensors data storage and (e) Data analysis.

### 2. 1 Libelium Sensors

Libelium Smart Cities Pro Model, along with particulate matter sensor (PM1, PM2.5, PM10), NO2, SO2, CO2, and the sensor for meteorological parameters (pressure, temperature, and humidity), have been used for data acquisition of TRAP as shown Fig. 1. The measured pollutants concentration was transmitted to Libelium Cloud Bridge and forwarded to AWS DynamoDB. The configuration was placed on a two-meter high stand, designed for sensors mounting. The study by Wu et al., (2002) shows that the maximum concentration of particulate matter occurs in-between heights of 2-3 meters. OPC-N3 is a particulate matter (PM1, PM2.5, and PM10) sensor by Libelium, which measures the concentration of particulate matter (Libelium, 2019). The air inhaled by the sensor passes through a built-in laser beam. Particles in the air scatter the laser light and determine the size and concentration of particles in return. Libelium has developed SO2-A4 sensor for SO2 measurement, “NO2-A43F” for NO2, “NE20-CO2 P-NCVSP” for CO2, and BME280 (Bosch Sensortec) for meteorological parameters (pressure, temperature, and humidity). Specification of all these sensors has been given in Table 1.

Libelium sensor probes for (a) particulate matter (PM1, PM2.5, PM10), (b) CO2, (c) NO2, (d) SO2, (e) sensor for meteorological parameters (pressure, temperature and humidity), (f) Libelium Smart Cities Pro Node, to which all sensor has been attached with six different sockets for various sensors.

Specifications of Libelium sensors.

### 2. 2 Sensors Deployment at Site

Outdoor TRAP was measured at the University of Peshawar (UOP) BRT Bus Station at coordinates 33.997795 N, 71.486272 E on February 24, 25, 27, and March 4, 9, 2020, as shown in Fig. 2(a) and Fig. 2(b). Indoor air quality is contaminated by pollutants that infiltrate from outside. Indoor air was monitored over 48 days from January 24, 2020, to March 12, 2020, for 24-hours, continuously, at the 3rd-floor University of Engineering and Technology Peshawar (main campus) at coordinates 34.001779 N, 71.485471 E. Also, the distance between the indoor and outdoor sites was approximately 450 m. Similarly, indoor measurements were taken at the height of 10 m, while outdoor at the height of 2 m. The sensor node and sensor probes were deployed at the designated location on the road for encountering maximum traffic emissions.

(a) TRAP measurement sites (source: Google Earth), (b) Deployed sensors at BRT.

### 2. 3 AWS for Storage and Analysis

This research opted for AWS cloud service for IoT data storage, analysis, and reporting. Data sent from the sensor to Libelium Cloud Bridge was then forwarded to AWS DynamoDB. With AWS IoT, sensors were connected through protocols for bi-directional communications (AWS, 2020). Sensor node was registered as a “Thing” in AWS IoT Core, a virtual representation of sensor in the cloud. Afterwards, a rule was established by simple query statements to extract the required data part from the sensor readings and forward it to Amazon DynamoDB. Further, the data was extracted from DynamoDB, transformed into tabular form for batch analysis and reporting, and loaded into IBM SPSS 25 and Excel.

### 2. 4 AQI Calculation

The daily average AQI for each pollutant was calculated from their concentration (Cp) based on US EPA procedures using Equation 1 (US-EPA, 2018) and US EPA breakpoints for AQI. The calculated AQI falls in one of the six predefined categories. Each category is represented by a colour scale that conveniently helps in knowing the level of hazard or toxicity of pollutants. The higher the AQI, the worse the air quality and vice versa.

 ${I}_{p}=\left(\frac{{I}_{high}-{I}_{low}}{B{P}_{high}-B{P}_{low}}\left({C}_{p}-B{P}_{low}\right)+{I}_{low}\right)$ (1)
Where,
Ip =the (air quality) index of a pollutant
Cp =the pollutant concentration,
BPlow=the concentration breakpoint that is ≤Cp
BPhigh=the concentration breakpoint that is ≥Cp
Ilow=the index breakpoint corresponding to BPlow
Ihigh=the index breakpoint corresponding to BPhigh

The values of BPlow, BPhigh, Ilow, and Ihigh can be found in US EPA breakpoints for AQI (US-EPA, 2018). For example, finding PM2.5’s AQI for a concentration of 40 μg m-3 (Cp=40 μg m-3) using Equation 1;

 ${I}_{P{M}_{2.5}=\left(\frac{150-101}{55.4-.35.5}\left(40-35.5\right)+101\right)=112}$

## 3. RESULTS AND ANALYSIS

Concentration and AQI analysis for indoor and outdoor condition has been discussed below. The upcoming analysis has given each pollutant concentration with its peak and minimum values.

### 3. 1 Outdoor Analysis

Outdoor data were measured on Feb 24-25, February 27, March 4, and March 9, 2020, with about 20 data points for each variable each day with 5 minutes of logging interval. Data was measured in the daytime with 90-minutes data recording on the first day, starting at 11:30 a.m. to 01:00 p.m. on February 24, another 120 minutes from 11:30 a.m. to 01:30 p.m. on February 25, 90 minutes from 12:00 p.m. to 1:30 p.m. on February 27, 120-minutes from 10:30 a.m. to 12:30 p.m. on March 4, while 60-minute data from 12:00 p.m. to 1:00 p.m. on March 9 was recorded. Outdoor data has been summarized in Table 2.

Descriptive statistics of outdoor measurements.

AQI between 101-150 corresponds to “Unhealthy for Sensitive Groups (USG).” The maximum AQI for outdoor pollutants was 169 (Unhealthy) for PM2.5, 203 (Very Unhealthy) for PM10, 181 (Unhealthy) for SO2, and 123 (Unhealthy) for NO2. SO2 contributes the largest to overall AQI as compared to other pollutants. From the analysis, four days overall outdoor AQI falls in the category of “Unhealthy,” and for a day, it was recorded as “Very Unhealthy.” Data analysis has shown that indoor AQI was slightly better than outdoor AQI. It was due to the distance and height of the indoor location, as mentioned in the upcoming comparison section. The unhealthy category has greater health consequences for the sensitive group (people with respiratory problems, heart diseases, the elderly, and children), while the general public is less affected. Further, according to US EPA, “for particle pollution, the sensitive groups include people with heart and lung disease, older adults, children, people with diabetes, and people of lower socioeconomic status (SES; a composite number indicating income, education, occupation).”

Vehicles were counted for five days in the daytime, as mentioned above. A total of 16,662 vehicles were counted for a duration of 1.5-hours on February 24, 18,916 for 2 hours, on February 25, 15,797 for 1.5-hours, on February 27, 20,156 on March 4 for 2 hours, and 11,211 for one hour on March 9, 2020. These traffic counts were used to find the correlation between measured pollutant concentration and the number of vehicles. Pearson correlation coefficient (r) shows that measured parameters are somewhat related to each other. A moderate positive correlation of 0.45 exists between motorcycles against PM1, and a low relationship of 0.27 exists between the former against NO2. Similarly, a correlation of 0.34 exists between humidity and PM10 while 0.47 between temperature and PM1. The rest of the correlations were not prominent, as those show little or no correlation. Outdoor measurement for PM2.5, PM10, SO2, NO2, and CO2 is plotted in the box and whisker plot in Fig. 3. The maximum daily mean for PM2.5 was 90 μg m-3 on February 27, while the maximum data spread was observed on March 4. Similarly, PM10 had the maximum daily mean of 356 μg m-3 on March 4 and the highest concentration of 1,074 μg m-3. Two outliers of 2,070 and 2,285 ppm for CO2 on February 25 have been removed, so boxplot for CO2 can be seen more apparently. Outdoor data has been recorded before the COVID-19 lockdown. The lockdown took place on March 13, 2020, in the city. The country observed smart lockdowns implemented in most affected areas, the frequency and degree (strictness) of which varies from place to place depending upon the severity of the disease.

Box and whisker plot for outdoor concentration (PM1, PM2.5, PM10, SO2, NO2, and CO2).

### 3. 2 Indoor Analysis

Indoor measurements were taken at the 3rd floor, New Academic Block, University of Engineering and Technology Peshawar (main campus). Data was recorded with 5-minute intervals, 24-hours continuously for 48 days, with 220 data points daily for each pollutant parameter each day starting from January 24, 2020, to March 12, 2020. Descriptive statistics of indoor data are given in Table 3.

Descriptive statistics of indoor measurements.

3. 2. 1 The Indoor Concentration of PM2.5, PM10, SO2, and NO2

Analysis of PM2.5 shows that the maximum daily average of 49 μg m-3 was noted on February 7, while the peak value of 294 μg m-3 was on February 19 and the minimum daily mean of 6 μg m-3 was noted on March 8 as shown in Fig. 4. The maximum daily mean for PM10 was 78 μg m-3 on February 19, while the minimum was 7 μg m-3 on March 8. Similarly, the concentration for SO2 was mostly below 200 ppb, as shown in Fig. 5, with a maximum daily mean of 465 ppb on January 24. Likewise, NO2 has the maximum daily mean of 247 ppb on February 1. The most considerable variation for NO2 was 437 ppb on March 6.

Boxplot showing the indoor concentration of PM2.5 and PM10.

Boxplot showing the indoor concentration of SO2 and NO2.

The highest concentration of particulate matter was recorded on February 19, as arranged in Table 4. The most probable cause could be weather variation (wind storm) on that day. Because of the scope of the current study, the effects of other factors such as wind speed/direction, temperature, rain, and humidity could not be included. The frequency percent for PM2.5 shows that range 1 to 50 μg m-3 is responsible for 90.4% of the total day distribution, as shown in Fig. 6. The rest, 9.6% of the distribution, is between 51-294 μg m-3, with only a 3% distribution above 200 μg m-3. Similarly, the highest concentration was 1,090 μg m-3 for PM10 on February 19. Frequency histogram shows that 1 to 50 μg m-3 is the highest recurring range with 67.6%, while 8.6% of the distribution is between 151 to 1,090 μg m-3, with merely a 6% distribution above 400 μg m-3. The PM10 AQI for February 19 was 62, which lay in the moderate range.

Descriptive statistics of indoor measurements on February 19, 2020.

Histogram frequency (%) of indoor concentration of PM2.5 and PM10 on Feb 19.

3. 2. 2 Indoor AQI of PM2.5, PM10, SO2, NO2, and Overall Indoor AQI

Indoor AQI for individual pollutants, PM2.5, PM10, SO2, NO2 is shown in Fig. 7. PM2.5’s AQI falls in three categorical limits. The first one, less than or equal to 50 AQI, corresponds to the “Good” category, consisting of 17% of the total data points measured during 48 days. Second, between 51 to 100 AQI corresponds to the “Moderate” category, making 52% of total data recorded, suggesting that PM2.5 belongs to the “Moderate” category for slightly more than half of the time. Lastly, the rest of 31% of the time, PM2.5’s AQI remains between 101 and 140 AQI, which corresponds to the “USG” category. Moreover, the maximum AQI of 135 for PM2.5 was on February 7, while a minimum AQI of 26 was noted on March 8, 2020. AQI of PM10 shows that 85% of the AQI values lie below 50 indexes, while the rest of 15% belongs to the “Moderate” category. It can be inferred from data that AQI for PM10 belongs to the “Good” category most of the time. Further, the maximum AQI was 62 for PM10 on February 19, while the minimum of 7 on March 8, 2020. The highest AQI for SO2 was 254 on January 24. By analyzing the data for SO2, 25 days belong to the category of “USG,” making a total of 66% of the total measurement. Similarly, SO2 AQI was “Unhealthy” for nine days that is 24% of the data, while the rest of the four days AQI were in the “Very Unhealthy” category making it 10% of the total data. AQI of NO2 resonates between 125 to 128 indexes and remains the “USG” category all of the time. Highest AQI was 128 for NO2, while the lowest of 107 was observed.

Indoor AQI of PM2.5, PM10, SO2, and NO2.

Data of overall indoor AQI shows that nine days, which makes 19% of the total measurement, belongs to the “Unhealthy” category, while 35 days AQI making 73% of the total belong to “USG” and the rest of 4 days making 8% of the total days belongs to “Very Unhealthy” category. Maximum overall AQI was found on January 24 with 254 indexes, while minimum AQI 123 was observed on March 8, 2020. Likewise, SO2 contributes the largest to overall AQI.

3. 2. 3 Frequency Analysis of Indoor Data

Frequency distribution for PM2.5, PM10, SO2, and NO2 has been shown in Fig. 8. PM2.5 has the highest frequency of 45% for the concentration interval of 1 to 20 μg m-3, while only 3% was above 40 μg m-3 concentration. Likewise, PM10 has the highest frequency of 45% for the concentration interval of 1 to 25 μg m-3, while 35% distribution is in the range of 26 to 50 μg m-3 and merely 1% above 100 μg m-3. Percentage frequency of SO2 shows that the highest frequency of 47% for the concentration interval of 100 to 150 ppb was found, while 23% distribution was in the range of 151 to 200 ppb. Similarly, NO2 has the highest frequency of 78% for the concentration interval of 225 to 250 ppb.

Histogram frequency (%) distribution of indoor PM2.5, PM10, SO2, and NO2 concentration.

### 3. 3 Comparison of Indoor and Outdoor Measurements

Interior and exterior AQI has been compared. It has been noted that overall indoor AQI was slightly better as compared to outdoor AQI. The comparative analysis concludes that the concentration and AQI of outdoor measurement for PM1, PM2.5, and PM10 were higher than indoor data, as Table 5 shows. The most probable reason for the betterment of indoor AQI was the distance (450 m) and height (10 m) of indoor location from the outdoor site. Similarly, outdoor measurements were taken at the designated spot on the road, which encountered a more significant amount of traffic-related air pollution than the indoor one. Moreover, there was no clear pattern for SO2 and NO2 concentration; sometimes outdoor takes lead while other-time indoor takes the lead, and this continues to their AQI. Also, it is difficult to arrive at an absolute judgment that outdoor concentration and AQI will be higher than indoor (valid in the case of PM1, PM2.5, and PM10, humidity and temperature), as it may seem, but analysis reveals otherwise. It can be deducted that more data is needed to arrive at a satisfactory conclusion. Because of the pilot study, the research scope was limited, which means the research needs to be continued for a longer time.

Comparison of indoor (In) and outdoor (Out) pollutant’s concentration.

## 4. DISCUSSION

TRAP in Peshawar city at BRT Peshawar University Bus Station was measured to know the outdoor concentration. Five days of data were collected for each pollutant each day. Likewise, indoor infiltration and air quality were monitored for 48 days at the 3rd floor, New Academic Block, University of Engineering and Technology Peshawar. The Libelium Smart Cities Pro model and Libelium’s sensor probes were used to find traffic-related air pollution, including PM1, PM2.5, PM10, NO2, SO2, and CO2, with additional meteorological parameters (pressure, temperature, and humidity). The data was sent to Libelium Cloud Bridge, afterwards directed to AWS IoT and AWS DynamoDB for storage and analysis.

The maximum AQI for outdoor pollutants was 169 (Unhealthy) for PM2.5, 203 (Very Unhealthy) for PM10, 181 (Unhealthy) for SO2, and 123 (USG) for NO2. SO2 dominates overall outdoor AQI. For all of those five days of outdoor reading, all of the time, overall AQI exceeds the 150 mark, which means that outdoor AQI falls in the category of “Unhealthy.” A correlation of 0.45 exists between motorcycle and PM1, while a low relationship of 0.27 exists between motorcycle against NO2. This means that the concentration was high whenever the number of motorcycles were high. Possibly, it was due to the local manufactured inexpensive motorcycles. Likewise, a correlation of 0.34 for humidity against PM10 while 0.47 between temperature and PM1 was noticed. The rest of the correlations were not prominent for further consideration.

For indoor measurement, PM2.5 has AQI of 135 (USG), PM10 has AQI of 62 (Moderate), NO2 has AQI of 254 (Very Unhealthy), while SO2 have an AQI of 128 (USG). Analysis shows that SO2, among other traffic-related air pollutants, contributes the largest to overall AQI. The overall indoor AQI remains in the “unhealthy” category by 19% of its total time, 73% in “USG,” while merely 8% in the category of “very unhealthy.” The sensitive group, which includes people with asthma, children, and older adults, are at higher risk. They need to be careful, take preventive standard operating procedures (SOPs), and make routine checkups. SOPs are standard instructions that help in lowering the unhealthy effects of pollutants. The unhealthy category, which includes people with lung diseases, children, and older adults, is susceptible at the highest and should avoid long or heavy exposure. While for the “very unhealthy category,” everyone should avoid extended outdoor activities and heavy exposure. Outdoor PM2.5 and PM10 concentrations were higher by 162% (4-Mar), and 452% (24-Feb) compared to indoor concentrations, as shown in Table 5.

A report by World Air Quality Index for PM2.5 shows that the top 37 out of 40 most polluted cities globally are from South Asia (IQAir, 2020). India, Bangladesh, and Pakistan have one of the worse PM2.5’s AQI in the world. Average AQI by Index (2020) from January 24 to March 12, 2020 (online available) was recorded at (a) Temple of Heaven, Dongcheng, Beijing, (b) Bandra, Mumbai, India, (c) UET Peshawar (current study), (d) Dhaka US Consulate, Bangladesh (PM2.5 AQI only). AQI for PM2.5 in Table 6 shows that all mentioned countries have exceeded the WHO 24-hours mean (25 μg m-3; which corresponds to AQI of 78). Pakistan’s AQI for PM2.5 was relatively better by 23%, 65%, and 170% compared to China, India, and Bangladesh. However, Pakistan’s AQI for SO2 and NO2 was poor as compared to the same countries. The most likely reason could be the Sulphur content of 500 ppm in Diesel and 3% in Furnace oil. Pakistan uses Euro II fuel (Sulphur content 500 ppm) and sub-standard fuel and is shifting to Euro V standard (Sulphur content 10 ppm). Secondly, the National Transport Research Center (NTRC, 2019) estimates that the number of motorcycles/scooters has increased by 219% (6.66 million to 14.62 million) in just four years from 2015 to 2019. The 2-strokes engines of motorcycles/ scooters and rickshaws produce most emissions due to incomplete fuel burning (Ministry of Finance, 2019b). Further, the industrial estate (Hayatabad, Peshawar.) is about 7.5 km away from the measurement location.

AQI comparison of neighbouring countries ( January 24 to March 12, 2020) (Index, 2020).

It was evident from the indoor data comparison of the current study with Rehan et al., (2021) that the AQI of PM2.5, PM10, and NO2 after lockdown improved from 135 to 49, 62 to 16, and 254 to 115 compared with AQI before lockdown, as shown in Table 7. The highest concentration of particulate matter was recorded on February 19. The most probable cause could be weather variation (wind storm) on that day. Effects of vehicle’s velocity, wind speed, vehicle condition, fuel type, traffic regulations, and other factors were not counted in this research because the scope of the study was limited to measure TRAP on a pilot scale. In addition, the exact reason for the higher correlation of motorcycles with PM1 and NO2 is unknown; however, the introduction of locally manufactured inexpensive motorcycles into the fleet may be the utmost reason. Further, a multidimensional approach is required to encounter the threat of TRAP, the core of which is measuring and monitoring TRAP continuously.

Comparison of maximum concentration and AQI before and after lockdown (indoor location).

## 5. CONCLUSION

The pilot study concludes that TRAP directly affects air quality deterioration. Data analysis of the pollutant’s concentration shows that the SO2 is the largest contributor to overall AQI for outdoor TRAP and indoor AQI, while PM10 was least contributive in indoor conditions. The maximum overall outdoor AQI was 203 due to PM10, while the maximum overall indoor AQI was 254 due to SO2. Moreover, indoor air quality was slightly better than outdoor air quality. Compared to the US EPA standard for AQI, the frequency analysis of PM2.5 concentration shows that 23% of concentrations were in “Good,” 54% in “Moderate,” while about 17% in the “USG” category of all recorded data. Similarly, PM10 has about 83% in “Good” while 15.6% in the “Moderate” category. SO2 has 73.5% in “USG” while 17% in “Unhealthy,” and NO2 has 97.4% in the “USG” category. This level of AQI translates that the sensitive groups are at higher risk as compared to ordinary people. Further, the peak concentration of particulate matter was observed during the daytime from 12:30 p.m. to 2:00 p.m. After in-depth analysis, this research recommends establishing AQI monitoring stations for continuous TRAP monitoring in the city. By doing so, citizens may have a better sense of air quality and would take preventive SOPs.

## Acknowledgments

Appreciations to undergraduate students Danish Hussain, Mushahid, Abdur Rehman, and M. Asif for helping in outdoor data measurement. I would like to acknowledge the support of Tufail Ahmad, National Centre in Big Data and Cloud Computing (NCBC), and Dr. Khurram Shahzad Khattak, Department of Computer System Engineering (DCSE) in AWS deployment.

## References

• Ahmad, I., Khan, B., Asad, N., Mian, I.A., Jamil, M. (2019a) Traffic-related lead pollution in roadside soils and plants in Khyber Pakhtunkhwa, Pakistan: implications for human health. International Journal of Environmental Science and Technology, 16(12), 8015-8022. [https://doi.org/10.1007/s13762-019-02216-7]
• Ahmad, I., Khan, B., Khan, S., Rahman, Z.u., Khan, M.A., Gul, N. (2019b) Airborne PM10 and lead concentrations at selected traffic junctions in Khyber Pakhtunkhwa, Pakistan: Implications for human health. Atmospheric Pollution Research, 10(4), 1320-1325. [https://doi.org/10.1016/j.apr.2019.03.003]
• Anwar, M.N., Shabbir, M., Tahir, E., Iftikhar, M., Saif, H., Tahir, A., Murtaza, M.A., Khokhar, M.F., Rehan, M., Aghbashlo, M., Tabatabaei, M., Nizami, A.-S. (2021) Emerging challenges of air pollution and particulate matter in China, India, and Pakistan and mitigating solutions. Journal of Hazardous Materials, 416, 125851. [https://doi.org/10.1016/j.jhazmat.2021.125851]
• Asseng, S., Ewert, F., Martre, P., Rötter, R.P., Lobell, D.B., Cammarano, D., Kimball, B.A., Ottman, M.J., Wall, G.W., White, J.W., Reynolds, M.P., Alderman, P.D., Prasad, P.V.V., Aggarwal, P.K., Anothai, J., Basso, B., Biernath, C., Challinor, A.J., De Sanctis, G., Doltra, J., Fereres, E., Garcia-Vila, M., Gayler, S., Hoogenboom, G., Hunt, L.A., Izaurralde, R.C., Jabloun, M., Jones, C.D., Kersebaum, K.C., Koehler, A.-K., Müller, C., Naresh Kumar, S., Nendel, C., O’Leary, G., Olesen, J.E., Palosuo, T., Priesack, E., Eyshi Rezaei, E., Ruane, A.C., Semenov, M.A., Shcherbak, I., Stöckle, C., Stratonovitch, P., Streck, T., Supit, I., Tao, F., Thorburn, P.J., Waha, K., Wang, E., Wallach, D., Wolf, J., Zhao, Z., Zhu, Y. (2014) Rising temperatures reduce global wheat production. Nature Climate Change, 5(2), 143-147. [https://doi.org/10.1038/nclimate2470]
• AWS. (2020). AWS Documentation: Internet of Things (IoT). Retrieved from https://docs.aws.amazon.com/index.html
• Berkowicz, R., Winther, M., Ketzel, M. (2006) Traffic pollution modelling and emission data. Environmental Modelling & Software, 21(4), 454-460. [https://doi.org/10.1016/j.envsoft.2004.06.013]
• Cesaroni, G., Boogaard, H., Jonkers, S., Porta, D., Badaloni, C., Cattani, G., Forastiere, F., Hoek, G. (2012) Health benefits of traffic-related air pollution reduction in different socioeconomic groups: the effect of low-emission zoning in Rome. Occupational and Environmental Medicine, 69(2), 133-139. [https://doi.org/10.1136/oem.2010.063750]
• Costa, L.G., Cole, T.B., Coburn, J., Chang, Y.C., Dao, K., Roque, P. (2014) Neurotoxicants are in the air: convergence of human, animal, and in vitro studies on the effects of air pollution on the brain. BioMed Research International, 2014, 736385. [https://doi.org/10.1155/2014/736385]
• Fontes, T., Li, P., Barros, N., Zhao, P. (2018) A proposed methodology for impact assessment of air quality traffic-related measures: The case of PM2.5 in Beijing. Environmental Pollution, 239, 818-828. [https://doi.org/10.1016/j.envpol.2018.04.061]
• Google, E. (Cartographer) (Jul 05, 2021) Google Earth photo of University of Engineering and Technology. Retrieved from https://goo.gl/maps/FmsNRd5rfZsjzs1e7
• Huang, L., Chen, M., Hu, J. (2016) Twelve-Year Trends of PM10 and Visibility in the Hefei Metropolitan Area of China. Advances in Meteorology, 2016, 1-9. [https://doi.org/10.1155/2016/4810796]
• Index, W.A.Q. (2020) Air Pollution in China, India, Pakistan and Bangladesh: Real-time Air Quality Index Visual Map. Retrieved from: https://aqicn.org/map/china, /
• IQAir. (2020) 2020 World Air Quality Report - Region and Cities PM2.5 Ranking. Retrieved from https://www.iqair.com/world-air-quality-report
• Kesavachandran, C.N., Kamal, R., Bihari, V., Pathak, M.K., Singh, A. (2015) Particulate matter in ambient air and its association with alterations in lung functions and respiratory health problems among outdoor exercisers in National Capital Region, India. Atmospheric Pollution Research, 6(4), 618-625. [https://doi.org/10.5094/apr.2015.070]
• Khan, M., Yasmin, T. (2014) Development of natural gas as a vehicular fuel in Pakistan: Issues and prospects. Journal of Natural Gas Science and Engineering, 17, 99-109. [https://doi.org/10.1016/j.jngse.2014.01.006]
• Kim, J.J., Smorodinsky, S., Lipsett, M., Singer, B.C., Hodgson, A.T., Ostro, B. (2004) Traffic-related air pollution near busy roads: the East Bay Children’s Respiratory Health Study. American Journal of Respiratory and Critical Care Medicine, 170(5), 520-526. [https://doi.org/10.1164/rccm.200403-281OC]
• Lelieveld, J., Klingmüller, K., Pozzer, A., Burnett, R.T., Haines, A., Ramanathan, V. (2019) Effects of fossil fuel and total anthropogenic emission removal on public health and climate. Proceedings of the National academy of Sciences, 116(15), 7192. [https://doi.org/10.1073/pnas.1819989116]
• Libelium. (2019) Smart Gases Pro Technical Guide http://www.libelium.com/downloads/documentation/gases_sensor_board_pro.pdf
• Lin, Y., Zou, J., Yang, W., Li, C.Q. (2018) A Review of Recent Advances in Research on PM(2.5) in China. International Journal of Environmental Research and Public Health, 15(3). [https://doi.org/10.3390/ijerph15030438]
• Linares, C., Díaz, J. (2010) Short-term effect of concentrations of fine particulate matter on hospital admissions due to cardiovascular and respiratory causes among the over-75 age group in Madrid, Spain. Public Health, 124(1), 28-36. [https://doi.org/10.1016/j.puhe.2009.11.007]
• Ministry of Finance, G.o.P. (2019a) Pakistan Economic Survey 2018-2019. Retrieved from https://www.finance.gov.pk/survey/chapters_19/Economic_Survey_2018_19.pdf
• Ministry of Finance, G.o.P. (2019b) Pakistan Economic Survey 2018-2019. Retrieved from https://www.finance.gov.pk/survey/chapter_11/16-Environment.pdf
• Montori, F., Bedogni, L., Bononi, L. (2018) A Collaborative Internet of Things Architecture for Smart Cities and Environmental Monitoring. IEEE Internet of Things Journal, 5(2), 592-605. [https://doi.org/10.1109/jiot.2017.2720855]
• Newman, N.C., Ryan, P.H., Huang, B., Beck, A.F., Sauers, H.S., Kahn, R.S. (2014) Traffic-Related Air Pollution and Asthma Hospital Readmission in Children: A Longitudinal Cohort Study. The Journal of Pediatrics, 164(6), 1396-1402.e1391. [https://doi.org/10.1016/j.jpeds.2014.02.017]
• NTRC (2019) Summary of Motor Vehicles on Road in Pakistan. Retrieved from https://www.ntrc.gov.pk/pakistan-2/ (accessed 15 Nov. 2021).
• Pascal, M., Corso, M., Chanel, O., Declercq, C., Badaloni, C., Cesaroni, G., Henschel, S., Meister, K., Haluza, D., Martin-Olmedo, P., Medina, S. (2013) Assessing the public health impacts of urban air pollution in 25 European cities: Results of the Aphekom project. Science of the Total Environment, 449, 390-400. [https://doi.org/10.1016/j.scitotenv.2013.01.077]
• Pérez-Martínez, P.J., de Fátima Andrade, M., de Miranda, R.M. (2015) Traffic-related air quality trends in São Paulo, Brazil. Journal of Geophysical Research: Atmospheres, 120(12), 6290-6304. [https://doi.org/10.1002/2014jd022812]
• Rehan, S., Sadiq, S., Ibraheem, M., Ahmad, A., Irfan, M.A. (2021) A Pilot Study of Indoor Air Quality at UET Peshawar. Paper presented at the 3rd International Conference on Sustainable Energy Technologies (ICSET 2021) Peshawar, Pakistan. https://uspcase.uetpeshawar.edu.pk/download/icset-2021-paper-proceedings/?wpdmdl=1461&refresh=61aa438a041711638548362
• Sánchez-Triana, E., Enriquez, S., Afzal, J., Nakagawa, A., Khan, A.S. (2014) Cleaning Pakistan's Air: Policy Options to Address the Cost of Outdoor Air Pollution. Washington, D.C., The World Bank. pp. 71-101.
• Shah, J., Mishra, B. (2016) IoT enabled environmental monitoring system for smart cities. Paper presented at the 2016 International Conference on Internet of Things and Applications (IOTA).
• US-EPA. (2006) How Air Pollution Affects the View. In U. S. E. P. Agency (Ed.), (pp. 1-2): Offce of Air Quality Planning and Standards.
• US-EPA. (2018) Technical Assistance Document for the Reporting of Daily Air Quality - the Air Quality Index (AQI). In EPA 454/B-18-007 September 2018: Airnow.com, .
• Wu, Y., Hao, J., Fu, L., Wang, Z., Tang, U. (2002) Vertical and horizontal profiles of airborne particulate matter near major roads in Macao, China. Atmospheric Environment, 36(31), 4907-4918. [https://doi.org/10.1016/S1352-2310(02)00467-3]
• Zanella, A., Bui, N., Castellani, A., Vangelista, L., Zorzi, M. (2014) Internet of Things for Smart Cities. IEEE Internet of Things Journal, 1(1), 22-32. [https://doi.org/10.1109/jiot.2014.2306328]
• Zhao, H., Che, H., Zhang, X., Ma, Y., Wang, Y., Wang, H., Wang, Y. (2013) Characteristics of visibility and particulate matter (PM) in an urban area of Northeast China. Atmospheric Pollution Research, 4(4), 427-434. [https://doi.org/10.5094/apr.2013.049]

### Fig. 1.

Libelium sensor probes for (a) particulate matter (PM1, PM2.5, PM10), (b) CO2, (c) NO2, (d) SO2, (e) sensor for meteorological parameters (pressure, temperature and humidity), (f) Libelium Smart Cities Pro Node, to which all sensor has been attached with six different sockets for various sensors.

### Fig. 2.

(a) TRAP measurement sites (source: Google Earth), (b) Deployed sensors at BRT.

### Fig. 3.

Box and whisker plot for outdoor concentration (PM1, PM2.5, PM10, SO2, NO2, and CO2).

### Fig. 4.

Boxplot showing the indoor concentration of PM2.5 and PM10.

### Fig. 5.

Boxplot showing the indoor concentration of SO2 and NO2.

### Fig. 6.

Histogram frequency (%) of indoor concentration of PM2.5 and PM10 on Feb 19.

### Fig. 7.

Indoor AQI of PM2.5, PM10, SO2, and NO2.

### Fig. 8.

Histogram frequency (%) distribution of indoor PM2.5, PM10, SO2, and NO2 concentration.

### Table 1.

Specifications of Libelium sensors.

Sensor Measuring range Accuracy Response time Max. power consumption
Particulate matter
(PM1, PM2.5, PM10)
Max. 2,000 μg m-3 at rate
up to 10,000 particles/second
from 0.35 μm to 40 μm
- - 270 mA @ 5 V
Nitric dioxide (NO2) 0 to 200 ppm ±0.2 ppm ≤30 seconds 1 mA
Sulfur dioxide (SO2) 0 to 20 ppm ±0.2 ppm and sensitivity of
320 to 480 nA/ppm
≤45 seconds 1 mA
Carbon dioxide (CO2) 0 to 5,000 ppm ±50 ppm for (0 to 2,500 ppm)
±200 ppm for (2,500 to 5,000 ppm)
≤60 seconds 80 mA
Temperature -40°C to +85°C ±1°C 1.62 seconds 1 μA
Relative humidity (RH) 0 to 100% <±3% 1 second 2.8 μA
Pressure 30 kPa to 110 kPa ±0.1 kPa - 4.2 μA

### Table 2.

Descriptive statistics of outdoor measurements.

Parameters PM1
(μg m-3)
PM2.5
(μg m-3)
PM10
(μg m-3)
CO2
(ppm)
SO2
(ppb)
NO2
(ppb)
Temp.
(°C)
RH
(%)
Mean 28 69 249 405 232 212 24 40
Maximum 61 167 1,074 2,285 373 252 29 54
Minimum 6 20 52 311 107 147 20 28
Median 26 62 201 359 232 214 24 40
N 100 100 100 100 81 100 100 100
Variance 146 761 32,490 67,588 3,481 423 7 60
Std. Dev 12 28 180 260 59 21 3 8
Range 55 147 1,022 1,974 265 105 9 26

### Table 3.

Descriptive statistics of indoor measurements.

Parameters PM1
(μg m-3)
PM2.5
(μg m-3)
PM10
(μg m-3)
CO2
(ppm)
SO2
(ppb)
NO2
(ppb)
Temp.
(°C)
RH
(%)
Mean 16 25 34 418 169 233 17 50
Maximum 49 294 1,090 2,737 560 437 28 58
Minimum 1 2 2 272 0 0 13 34
Median 15 23 28 349 142 240 17 51
N 10,584 10,582 10,581 10,585 7,450 7,314 10,583 10,584
Variance 87 282 1,094 54,684 8,293 1,678 7 20
Std. Dev 9 17 33 234 91 41 3 5
Range 48 292 1,088 2,466 560 437 15 25

### Table 4.

Descriptive statistics of indoor measurements on February 19, 2020.

Parameters PM1
(μg m-3)
PM2.5
(μg m-3)
PM10
(μg m-3)
CO2
(ppm)
NO2
(ppb)
Temp.
(°C)
RH
(%)
Mean 16 38 78 698 229 18 18
Maximum 45 294 1,090 2,128 250 21 21
Minimum 2 3 3 426 155 18 18
N 210 210 210 210 210 210 210
Variance 100 2,238 24,591 119,515 149 0 0
Std. Dev 10 47 157 346 12 1 1
Range 43 291 1,087 1,702 95 3 3

### Table 5.

Comparison of indoor (In) and outdoor (Out) pollutant’s concentration.

Date PM1 (μg m-3) PM2.5 (μg m-3) PM10 (μg m-3) CO2 (ppm) SO2 (ppb) NO2 (ppb)
In Out In Out In Out In Out In Out In Out
24-Feb 22 26 36 70 46 254 345 361 223 226 228 217
25-Feb 19 28 29 55 35 132 493 502 220 223 226 210
27-Feb 24 47 46 90 81 229 517 406 271 258 269 209
04-Mar 12 22 28 74 99 356 450 368 238 221 160 219
09-Mar 15 12 38 54 92 284 381 352 235 230 142 196

### Table 6.

AQI comparison of neighbouring countries ( January 24 to March 12, 2020) (Index, 2020).

Country PM2.5 PM10 SO2 NO2 Overall AQI
China (Beijing) 101 67 7 17 101
India (Mumbai) 136 39 1 12 136
Pakistan (Peshawar - current study) 82 32 153 125 153
Bangladesh (Dhaka) 221 - - - 221

### Table 7.

Comparison of maximum concentration and AQI before and after lockdown (indoor location).

Parameters PM2.5 (μg m-3) PM10 (μg m-3) SO2 (ppb) NO2 (ppb)
Before After Before After Before After Before After
Concentration (Maximum) 49 11.7 78 17.6 247 173 465 177
AQI 135 49 62 16 128 145 254 115