Characterization of roadside air pollutants: An artery road of Lanzhou city in northwest China

. Due to the lack of air and traffic monitoring data in China, this study employed a remote sensing (RS) system and air quality monitoring sensors to investigate the air pollutant concentrations, traffic profiles and meteorological data in order to characterize roadside pollutant concentrations and identify their influence factors for a major road in Lanzhou. The results suggested that the temporal variations of concentrations of SO 2 , NO 2 , PM 2.5 , PM 10 and CO peaked at 5:00-7:00 while O 3 peaked at 17:00. The concentrations of roadside air pollutants were lower on weekdays than weekends except O 3 . The concentrations of PM 2.5 and PM 10 declined with the height of the monitoring sites increased, peaking at 2-meter height whereas SO 2 , NO 2 , CO, and O 3 reached their highest concentrations at 4-meter height. The concentrations of roadside NO 2 and CO show positive correlations with the volume of heavy-duty trucks (HDTs) and light-duty passenger vehicles (LDPVs), respectively. This study provided scientific support to the establishment of roadside air quality monitoring systems and the assessment of the environmental and health impact from vehicle emissions.


Introduction
With the growing motor vehicle ownership in China, exhaust emissions have become increasingly prominent in urban and regional air pollution [1][2][3][4] .Statistics show that China registered 372 million motor vehicles and there were 15.93 million tons of emissions from four vehicle-based pollutants at the end of 2020, where mobile sources were the primary contributor to PM2.5 in certain cities [5] .Several studies have shown that vehicular NOX is the culprit of high NO2 concentrations in urban areas.In addition, VOCs and NOX emissions from vehicles are important precursors of O3, contributing significantly to summertime O3 pollution [6][7][8] .Meanwhile, such emissions, being near the earth's surface, have a particular impact on the health of populations subject to long-term exposure to traffic pollution or high traffic density [9][10][11] .
There have been few studies on the characteristics of air pollution at the level of traffic monitoring sensors in China, and the existing ones mainly focus on large cities such as Beijing, Shanghai, Guangzhou, and Shenzhen [12][13][14] .The 14th Five-Year Plan for Ecological Environment Monitoring clearly states that roadside air quality sensors shall be built along major arterial roads in municipalities for the monitoring of PM2.5, NMHC, NOX and traffic flow.Lanzhou, the capital of the Gansu province at the geometric center of China, is an important hub for transportation and a golden node of the Belt and Road Initiative.As the first victim of photochemical smog in China, Lanzhou has seen its vehicle ownership soaring from 89.9 in 2016 to 1.145 million in 2020, growing 6.25% per year, higher than the provincial average of 5.2%.Compounded by factors such as its industrial, geographical and climate features, Lanzhou paints a bleak picture of atmospheric pollution [15] .A wealth of studies has demonstrated that Lanzhou are on an upward trend in terms of NO2 and O3 concentrations [16][17][18] .
This paper investigates the spatiotemporal variations in roadside air pollutant concentrations and analyzes their influencing factors with Lanzhou as the case study, where pollutant concentrations, meteorological data, and traffic flow were watched by the roadside monitoring sensors and vehicle exhaust RS (remote sensing), aiming to provide scientific evidence to support the design and deployment of traffic-based air quality monitoring system, environment impact assessment of vehicle emissions, and relevant research on public health.

Road selection
In 2016, to enhance the prevention and control of vehicle exhaust pollution, fleet screening for high emitters, and air quality in the city, Lanzhou's Bureau of Ecology and Environment selected six sites along the entrances and exits of urban areas and major transport linkages to install RS for the monitoring of vehicle emissions.
In consideration of traffic flow, site conditions and other factors, the study selected the Fuci Street in Chengguan District as the section during the monitoring period of July 1-30, 2021.Fuci Street, as the artery of National Highway G109, is a key passage linking north and south Lanzhou.

Monitoring sites
In accordance with the topographic and geomorphological features of the selected 1km-long road section, researchers used pole-mounted units and set up four air quality monitoring sites spaced 500 meters apart and distributed symmetrically on both sides of the road, with each monitoring site fitted with monitoring equipment at a height of 2m, 4m and 6m off the ground.See Table 1 for the longitude and latitude information of the four monitoring sites, and Figure 1 for the roadside monitoring system deployment.RS for vehicle exhaust monitoring is also available on the selected road.

Monitoring equipment
A total of 12 air quality monitoring sensors and 1 vehicle exhaust remote sensor were used in this study.The 12 devices are XHAQSN-812 Air Quality Sensor (Micro-station) from Sailhero Environmental Protection, which monitor the six conventional air pollutants: SO2, NO2, PM2.5, PM10, CO, and O3, as well as meteorological parameters including wind speed, wind direction, temperature, humidity, and atmospheric pressure.The exhaust sensor is a Hikvision iDS-TCV300 License Plate Recognition (LPR) unit by Sailhero's XHRCM2000 Continuous Particulate Monitor.
Consisting of high-precision electrochemical and optical sensors, the solar-powered air quality sensors, with low detection limits, reliable data, high time resolution, compact size among other merits, are suitable for grid and dense layout, which have numerous monitoring applications including air quality in cities and industrial parks, as well as vehicle exhausts in the roadside atmosphere.3 Results and analysis

Characteristics: concentrations of roadside air pollutants
From the monitored data of atmospheric pollutants by hour from July 1st to 30th, this study conducts a statistical analysis on the variations by day, by weekday/weekend, and by height.
In general, SO2, NO2, PM2.5, PM10, and CO pollution levels were 5.2%, 10.3%, 1.5%, 6.7%, and 8.2% lower on weekdays than weekends, while the weekend O3 pollution was 3.3% lower.Specifically, SO2 and CO were higher on Tuesday and lower on Monday, NO2 and PM10 higher on Saturday and lower on Monday and Wednesday, respectively, PM2.5 higher on Thursday and lower on Friday, while O3 higher on Friday and lower on Sunday.Figure 4 shows the mass concentrations variations of the six pollutants at different levels above ground.Both PM2.5 and PM10 gradually decline with height, with the average concentrations at 2-meter height being 23.9μg/m 3 and 64.0μg/m 3 , and 6-meter height 22.1μg/m 3 and 61.2μg/m 3 , which were 7.8% and 4.4% lower compared with 2m height.SO2, NO2, CO, and O3 first increased before decreasing with the highest concentrations at 4-meter height and the lowest at 2-meter height.The average concentrations at 4-meter height were 6.7μg/m 3 , 54.6μg/m 3 , 0.55mg/m 3 and 103.7μg/m 3 , and the 2-meter average 5.1μg/m 3 , 43.2μg/m 3 , 0.50mg/m 3 , and 95.0μg/m 3 .Compared with 4-meter level, the concentrations of SO2, NO2, CO and O3 at 2-meter was lower by 23.1%, 20.9%, 9.7% and 8.4%, respectively.The vertical variations in pollutant concentrations may be related to the E3S Web of Conferences 360, 01039 (2022)  https://doi.org/10.1051/e3sconf/202236001039VESEP2022 location of vehicle exhausts, physicochemical properties of pollution, as well as diffusion and dispersion processes.

Influencing factor analysis: concentrations of roadside atmospheric pollutants
Some studies [8,19] have found that NOX is mainly produced by vehicle exhausts and CO by coal combustion and vehicle exhausts.Figure 5 shows the correlation between NO2 and CO mass concentrations and roadside traffic flow.NO2 and CO concentrations were largely influenced by traffic flow, showing a positive correlation with the correlation coefficients of NO2 being 0.72 and 0.94, while of CO being 0.78 and 0.62 for day 1-5 and day 25-30, respectively.In addition, NO2 and CO exhibited similar trends in daily variations, where the correlation coefficient reached as high as 0.81.
Statistics [5] show that NOX emissions from diesel vehicles and CO from gasoline vehicles both exceed 80% of total vehicle emissions.The study selected the traffic volume of HDTs (heavy duty trucks) and LDPVs (light duty passenger vehicles) to analyze their correlation with NO2 and CO.As shown in Figure 5, the concentrations of roadside NO2 is mainly pegged to the volume of HDTs while CO to LDPVs.The correlation coefficients of roadside NO2 concentrations and HDTs were 0.94 and 0.84, and CO to LDPVs 0.79 and 0.60 in the representative time periods of day 1-5 and day 25-30.Meteorological conditions also play a key role in the concentrations of air pollutants on urban roads.The 30-day monitoring Fuci Street shows that the monthly average wind speed, relative humidity, and temperature were 0.87m/s, 36.5%, and 27.9°C, respectively.Figure 6 is the correlation between NO2, PM2.5, CO, O3 concentrations and meteorological parameters in the roadside atmosphere.The concentrations of NO2, PM2.5, and CO was negatively correlated with wind speed, while the correlation coefficient between PM2.5 and wind speed was -0.60.The greater the wind speed, the more dispersion, and the lower roadside concentrations of NO2, PM2.5, and CO.The higher the wind speed, the less dispersion, and the higher roadside concentrations of NO2, PM2.5, and CO.Roadside O3 is mainly produced as a result of a complex set of photochemical reactions that involve precursors NOX and volatile organic compounds (VOCs).NO2, as a precursor to ground-level O3, is consumed when weather conditions are favorable for O3 formation, hence NOX and O3 display opposite trends.
In general, higher relative humidity leads to poorer dispersion, causing pollutants to concentrate in the roadside atmosphere.There is negative correlation between the concentrations of NO2, PM2.5, CO, O3 and relative humidity with the correlation coefficient between NO2 and relative humidity being -0.67, which may be related to the summer monitoring period when it's hotter and drier in Lanzhou with an average relative humidity of 18-56%.Roadside O3 concentrations is closely linked to meteorological elements such as temperature, humidity, radiation, and boundary layer height.Heat, humidity, and light are conducive to O3 pollution.
The average wind speed was only 0.87m/s during the monitoring period, determined mainly by the geographical location and climatic features of Lanzhou.Studies [20] show that the northeast trade winds are dominant in Lanzhou with the average wind speed being only 0.9m/s.Due to the meteorological conditions including weak winds, downward pollutant flow, strong atmospheric stability, unfavorable dispersion conditions, Lanzhou suffers from stagnant air, trapping and accumulating local pollution.Motor vehicles are a local source of air pollutants, therefore, reducing vehicle emissions can improve the air quality in Lanzhou.

Conclusion
Based on the new policies on ecological monitoring in the 14th Five-Year Plan and a lack of air and traffic monitoring data, this study characterized roadside air pollutant concentrations and its influencing factors for a major road in Lanzhou, where a remote sensing (RS) system and air quality monitoring sensors were employed to investigate roadside air pollutant concentrations, traffic profiles, and meteorological data.The major findings can be summarized as follows: (1) The temporal variations of concentrations of SO2, NO2, PM2.5, PM10 and CO peaked at 5:00-9:00 while O3 peaked at 15:00.The concentrations of SO2, NO2, PM2.5, PM10 and CO during the peak periods were 2.9, 3.6, 1.8, 2.3, 2.9 and 3.1 times higher than those during the off-peak periods, respectively.In addition, SO2, NO2, PM2.5, PM10, and CO concentrations on weekdays were 5.2%, 10.3%, 1.5%, 6.7%, and 8.2% lower than weekends.The concentrations of PM2.5 and PM10 declined with the height of the monitoring sites increased, peaking at 2-meter height whereas SO2, NO2, CO, and O3 first reached their highest concentrations at 4-meter height.
(2) Roadside NO2 and CO concentrations showed positive correlations with the traffic volumes of HDTs and LDPVs, respectively.The concentrations of NO2, PM2.5, and CO are negatively correlated with wind speed, and there is negative correlation between the concentrations of NO2, PM2.5, CO, O3 and relative humidity.
The equipment selection is mainly based on the Technical Specification for Installation and Acceptance and Operation of Grid-based Monitoring System for Air Pollution Prevention and Control (DB13/T2546-2017) and the Technical Specification for Automatic Ambient Air Quality Monitoring (HJ193-2005).
This study was supported by the Ministry of Transport (The research on the methodology of on-road vehicle emission calculation), Transport Planning and Research Institute Ministry of Transport (The research on the technology of the inspection and maintenance of heavy-duty trucks), the National Development and Reform Commission (The control of the environmental pollution from the regional coal freight by trucks) and Laboratory of Transport Pollution Control and Monitoring Technology.

Table 1 .
Longitude and latitude information of the monitoring sites.