Impact of the covid-19 event on trip intensity and air quality in southern china

The COVID-19 epidemic discovered and reported at the end of December 2019 and began spreading rapidly around the world. The impact of the COVID-19 event on the trip intensity, AQI (air quality index), and air pollutants, including PM2.5, PM10, SO2, CO, NO2, and O3in Shenzhen, Guangzhou, and Foshan (the so-called ‘three cities’) from January 12 to March 27, in 2019 and 2020, are compared and discussed. In 2020, the combined trip intensity in the three cities ranged between 0.73 and 5.54 and averaged 2.57, which was 28.4% lower than that in 2019. In terms of the combined AQIs for the three cities, from January 12 to March 26, 2020, the daily AQIs ranged between 21.0 and 121.3 and averaged 56.4, which was 16.0% lower than that in 2019. The average AQIs in order were Guangzhou (57.5) &gt; Foshan (54.1) &gt; Shenzhen (44.1). In 2019, the distribution proportions of the six AQI classes were 45.2%, 50.4%, 4.40%, 0%, 0%, and 0%, respectively, while those in 2020 were 62.7%, 37.3%, 0%, 0%, 0% and 0%, respectively. For the combined data for the three cities, on the top five days with the highest AQIs during the epidemic period, the average concentrations of PM2.5, PM10, SO2, CO, NO2, and O3were 76.4 µg m–3, 113.4 µg m–3, 5.14 ppb, 0.88 ppm, 36.5 ppb and 55.5 ppb, which were 55.2%, 49.4%, 55.1%, 30.0%, 45.1% and 15.5% lower than those during the non-epidemic period (from January 12 to March 27, 2017–2019). The above results revealed that the comprehensive strict epidemic prevention and control actions reduced trip intensity and improved the air quality significantly.


INTRODUCTION
Since the economic reform in 1978, China's economy has been in a state of rapid development (Chan and Yao, 2008). China has become an important driving force for the world's economic development. China's rapid economic development is accompanied by increased energy demand and health problems caused by air pollution.
A series of consequences caused by air pollution problems have attracted the attention and research of all countries in the world. A large number of research results have proved the obvious impact of air pollution on people's health (Pope and Dochery, 2006;Cao et al., 2012;Heal et al., 2012;Pope and Dochery, 2013;Jin et al., 2017). In 1952, there was a two-day sulphate environmental incident in London that eventually killed about 12,000 people. Hu et al. (2015) studied the impact of PM 2.5 on human health, pointing out that fine particles can increase the incidence of respiratory and cardiovascular diseases. PM 10 levels have a significant impact on mortality from cardiovascular and respiratory diseases (Samet et al., 2000;Liang et al., 2016;Wang et al., 2018). Sulfur dioxide in the air accumulates in the body through human respiration and affects the respiratory system, causing respiratory diseases such as asthma (Xu et al., 2020b). It is a cause for concern because it plays an important role in the formation of ozone, particle pollution, smog and acid rain (Kyrkilis et al., 2007). The impact of air pollution on human health and the human living environment makes air quality a serious issue for the Chinese government and the public (Xu et al., 2020a;Zhang et al., 2020).
On December 31, 2019, the WHO issued a warning that an unknown infectious disease was spreading (Atri et al., 2020). On February 11, 2020, WHO officially named this epidemic as COVID-19 and explained that this epidemic

Air Quality Index (AQI)
The sub-AQI of the six criteria pollutants were first calculated with the observation concentrations, as shown in Eq. (1) (She et al., 2017;Shen et al., 2017). The overall AQI represents the maximum of the sub-AQI of all pollutants, where when the AQI is higher than 50, the highest sub-AQI contributor is defined as the primary pollutant on that day, as shown in Eq. (2) (She et al., 2017;Shen et al., 2017): (1) AQI = max(I 1 , I 2 , …, I n ) (2) IAQI P : the air quality sub index for air pollutant p. C P : the concentration of pollutant p. C low : the concentration breakpoint that is ≤ C P . C high : the concentration breakpoint that is ≥ C P . I low : the index breakpoint corresponding to C low . I high : the index breakpoint corresponding to C high . The daily AQIs were calculated based on the 24-hour average concentrations of SO 2 , NO 2 , PM 2.5 , PM 10 , CO, and the daily average 8-hour maximum concentration of O 3 . From the United States Environmental Protection Agency (U.S. EPA) AQI, the ranges of AQI values related to air quality can be classified into six classes: Grade I: 0-50 (Good, Green), Grade II: 51-100 (Moderate, Yellow), Grade III: 101-150 (Unhealthy for Sensitive Groups; Orange), Grade IV: 151-200 (Unhealthy; Red), Grade V: 201-300 (Very unhealthy; Purple), and Grade VI: 300-500 (Hazardous; Maroon) (Hu et al., 2015;Lanzafame et al., 2015;She et al., 2017;Zhao et al., 2018).

Wind Streamlines and Wind Speeds
In order to understand the path of airflow in Guangdong Province from January to March in 2019 and 2020, respectively, we chose GrADS (Grid Analysis and Display System, http://cola.gmu.edu/grads/) to compute and draw the distribution of the monthly average near-surface streamlines and wind speed with NCEP GDAS/FNL 0.25 Degree Global Tropospheric Analyses data (https://rda.ucar.edu/dat asets/ds083.3/). During this period, the pathway of airflow is usually affected by the Siberia High (Siberia Anticyclone).

Comparison for Air Pollutants and Trip Intensity
The daily trip intensity in Shenzhen, Guangzhou and Foshan from January 12 to March 27, 2019 and 2020, are shown in Fig. 2. The daily trip intensity and daily concentrations for PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 from January 12 to March 27, 2019 and those for 2020, are shown and compared in Figs. 3(A)-3(F), respectively.

Trip Intensity
With the development of cloud database technology, the modeling and analysis of public big data plays an increasingly important role in epidemic warning, analysis and rapid response (Ajelli et al., 2010). In order to analyze and compare the changes in the trip intensity in Shenzhen, Guangzhou and Foshan during the epidemic control period, the daily trip intensity between January 12 and March 27, 2019 and those of 2020 for the three cities were shown in Fig. 2.
As shown in Fig. 2(a), in Shenzhen, from January 12 to March 27 in 2019, the trip intensity ranged between 0.94 and 4.61, and averaged 3.45; the top five highest daily averages ranged from 4.38 to 4.54 and averaged 4.48 (from March 10-March 14, 2019), while the lowest five daily averages were between 0.94-1.00 and averaged 0.97 (from February 4-February 8, 2019). However, in Shenzhen, from January 12 to March 26 in 2020, the trip intensity ranged between 0.73 and 4.97 and averaged 2.46, which was 28.8% lower than that in 2019. The top five highest daily average ranged from 4.46 to 4.97 and averaged 4.78 (from January 13-January 17, 2020), which was 6.7% higher than those in 2019. The lowest five daily average ranged between 0.73-0.78 and averaged 0.76 (from January 26-January 30, 2020), which was 21.7% lower than that in 2019. As shown in Fig. 2 (a), the trip intensity in Shenzhen remained stable around its lowest point from February 4 to February 8, 2019. However, that in 2020 fell between January 25 and February 9, which was due to the comprehensive strict epidemic prevention and control actions taken in Guangdong Province, such as shutting down factories and traffic control. These actions greatly reduced the trip frequency.
As shown in Fig. 2(b), in Guangzhou, from January 12 to March 27 in 2019, the trip intensity ranged between 1.47 and 4.61, and averaged 3.57. The top five highest daily averages ranged from 4.35 to 4.59 and averaged 4.49 (from March 10-March 14, 2019), while the lowest five daily averages ranged between 1.47-1.64 and averaged 1.53 (during February 4-February 8, 2019). However, in Guangzhou, from January 12 to March 26 in 2020, the trip intensity ranged between 0.93 and 5.54, and averaged 2.48, which was 30.1% lower than that in 2019; the top five highest daily average ranged from 4.17 to 5.54 and averaged 4.67 (from January 13-January 17, 2020), which was 4.0% higher than those in 2019. The lowest five daily average ranged between 0.93-0.98 and averaged 0.95 (from January 27-January 31, 2020), which was 37.9% lower than those was in 2019. As shown in Fig. 2(b), the trip intensity in Guangzhou remained stable around its lowest point from February 4 to February 9, 2019. However, the lowest trip intensity in 2020 was between January 26 and February 9. As was the case with Shenzhen, the local traffic trip frequency in Guangzhou also was reduced due to the impact of strict control measures during the epidemic period.
As shown in Fig. 2 from 4.20 to 4.67 and averaged 4.42 (from January 12-January 16, 2020), which was 6.4% lower than those in 2019. The lowest five daily averages ranged between 0.92-1.01 and averaged 0.96 (from January 27-January 31, 2020), which was 34.3% lower than those in 2019. As shown in Fig. 2(c), the trip intensity in Foshan remained stable around its lowest point from February 4 to February 9, 2019. However, the lowest trip intensity in 2020 was between January 27 and February 9, which was due to the restrictions measures resulting in a reduction in travel.
It can be seen that compared to the same period in 2019, the average trip intensity in the three cities decreased significantly. This indicated that strict prevention and control measures affected the travel plans of local residents. Also, the reduction in human travel activity may had impact on the atmospheric environment. Between January 12 and March 27, 2019 and the same period in 2020, the lowest values of trip intensity in the three cities all occurred during the Chinese New Year holiday. For the purpose of controlling the spread of COVID-19, China adopted self-quarantine for residents in the three cities, which caused the residents to cancel unnecessary trip plans, and in turn led the trip intensity to remain around its lowest point for about a week. After February 14, 2019 and 2020, the troughs in the trip intensity all occurred on a Sunday, which means that the impact of the holidays on people's travel activities was obvious. PM2.5 Concentration PM 2.5 is also known as fine particles, which refers to particles smaller than 2.5 micrometers in diameter. PM 2.5 comes from production and daily combustion, including fuel vehicles, thermal power station, natural fires, agricultural combustion and some industrial processes (Kyrkilis et al., 2007). As an important component of aerosol, PM 2.5 may affect the environment by changing the atmospheric radiation balance, reducing soil nutrient levels, lowering visibility, degrading ecosystem biodiversity and so on (Chen et al., 2015;Myhre, 2009). Hu et al. (2015) studied the impact of PM 2.5 on human health, pointing out that fine particles can induce cardiovascular and respiratory diseases in the body. The daily concentration of PM 2.5 and trip intensity from January 12 to March 27, 2019 and those of 2020 are presented in Fig. 3(A), respectively.
As shown in Fig. 3(A), between January 12 and March 27, 2019, in Shenzhen, Guangzhou, and Foshan, the PM 2.5 concentrations ranged between 11.0 and 74.0, between 9.00 and 92.0 and between 10.0 and 81.0 µg m -3 and averaged Fig. 3(A). The trip intensity and daily PM 2.5 concentrations from January 12 to March 27 2019 and those in 2020 during the same period in Shenzhen, Guangzhou, and Foshan, respectively. 25.9, 34.2, and 34.1 µg m -3 , respectively. Those during 2020 ranged from 5.00-40.0, 5.00-44.0, and 4.00-43.0 µg m -3 , and averaged 21.5, 23.6, and 22.9 µg m -3 , respectively. These values in 2020 were 16.8%, 31%, and 32.7% lower than those from January 12 to March 27, 2019. Based on the combined data from the three cities, from January 12 to March 26 2020, the average PM 2.5 decreased by 27.7% compared with that during the same period in 2019. When the trip intensity of Shenzhen, Guangzhou, and Foshan were less than 4.4, 4.3 and 3.3, respectively, the PM 2.5 concentrations corresponding to the same trip intensity were higher in 2020 than in 2019. Primary and secondary combustion products mainly come from motor vehicles and thermal power stations, which are two crucial sources of fine particles (PM 2.5 ). Therefore, this change was due to the fact that during the epidemic control period, while the trip intensity was reduced, the power plants continue normal operation and produced atmospheric pollutants.
It can be seen that compared with 2019, the concentration of PM 2.5 decreased significantly from January 12 to March 26, 2020. There may be two reasons for this decline. The first reason is that January 24 is the traditional Chinese New Year, so a large number of factories are on holiday. This greatly reduces the industrial sources of PM 2.5 . The second, news about the COVID-19 epidemic gradually spread, and people canceled visits to go outside during the first lunar month. After entering February, the Chinese government announced a number of home segregation measures. This reduced the PM 2.5 generation due to restrictions on transportation. In March, people took the initiative to avoid going out for self-protection, and factory operations resumed slowly, which means that the amounts of fine particles from motor vehicles and industrial production is greatly reduced. With the gradual reduction of industrial sources, the influence of trip intensity on PM 2.5 concentration changes was therefore relatively greater.

PM10 Concentration
PM 10 levels have a significant impact on mortality from cardiovascular and respiratory diseases (Samet et al., 2000;Liang et al., 2016;Wang et al., 2018). Researchers point out that atmospheric particulates carry acidic and toxic substances, including polycyclic aromatic hydrocarbons and heavy metals, which are believed to have an impact on human health (Song et al., 2008;Tao et al., 2009;Matus et al., 2012;Sun et al., 2014;Liang et al., 2016). The daily concentration of PM 10 and trip intensity from January 12 to March 27, 2019 and those for the same period in 2020 are presented in Fig. 3(B), respectively.
As shown in Fig. 3(B), between January 12 and March 27, 2019, in Shenzhen, Guangzhou, and Foshan, the PM 10 concentrations ranged between 18.0 and 119, between 15.0 and 124, and between 19.0 and 149 µg m -3 and averaged 40.9, 55.2, and 62.3 µg m -3 , respectively. While those during 2020 ranged from 9.00-65.0, 6.00-75.0, and 6.00-87.0 µg m -3 and averaged 35.2, 38.5, and 39.6 µg m -3 , respectively, which was 13.9%, 30.2%, and 36.5% lower than that from January 12 to March 27, 2019. Based on the data from the three cities, from January 12 to March 26 2020, the average PM 10 concentration decreased by 28.5% compared with that during the same period in 2019. When the trip intensity of Shenzhen, Guangzhou, and Foshan was less than 4.3, 4.2, and 3.2, respectively, the PM 10 concentrations corresponding to the same trip intensity were higher in 2020 than in 2019. Sources of coarse particulate (PM 10 ) matter mainly includes nature and human activities, such as sand and dust storms, natural dust and industrial production activities (Querol et al., 2004;Xu et al., 2017). When the concentration of particulate matter discharged by a motor vehicle is reduced due to reduction in trip intensity, movement of polluted air generated by heating from northern China to southern regions can significantly affect the local PM 10 levels in the three cities.
It's clear that PM 10 concentrations in the three cities from January 12 to March 26 in 2020 were greatly reduced, indicating that the reduction in particulate mass emissions was caused by factory shutdowns and people staying home during the Chinese New Year holiday and the epidemic control period.

SO2 Concentration
Sulfur dioxide (SO 2 ) is a colorless, reactive gas produced when burning sulfur-containing fuels such as coal and petroleum (Kyrkilis et al., 2007;Khattak et al., 2013). The daily concentration of SO 2 and trip intensity from January 12 to March 27 2019 and those in the same period in 2020 are presented in Fig. 3(C), respectively.
As shown in Fig. 3(C), between January 12 and March 27, 2019, in Shenzhen, Guangzhou and Foshan, the SO 2 concentrations were between 1.05 and 2.80 ppb, 1.40 and 3.50 ppb, and 1.05 and 5.60 ppb and averaged 1.71, 2.11, and 2.96 ppb, respectively. Those in 2020 ranged from 1.75-2.45 ppb, 1.40-3.15 ppb, and 1.40-3.50 ppb, and averaged 1.82, 2.02, and 2.14 ppb. The average value of SO 2 in Guangzhou and Foshan in 2020 decreased by 4.2% and 27.6% compared with that during the same period in 2019, while Shenzhen increased by 6.3%. Based on the combined data from the three cities, from January 12 to March 26 2020, the average SO 2 decreased by 11.8% compared with that during the same period in 2019. When the trip intensity of Shenzhen, Guangzhou and Foshan were less than 4.4, 4.5, and 2.6, respectively, the SO 2 concentrations corresponding to the same trip intensity were higher in 2020 than in 2019. Sources of sulfur dioxide (SO 2 ) mainly include human activities and natural processes, such as the burning of biofuels, motor vehicle emissions, animal catabolism, natural sulfur cycles, and volcanic eruption (Kettle and Andreae, 2000;Halmer et al., 2002;Dentener et al., 2006). When the concentration of SO 2 emitted by the fossil fuel combustion was reduced due to the reduction in trip intensity, the movement of air in northern China with excessive sulfur dioxide (SO 2 ) concentration to the three cities can affect the local sulfur dioxide level.
Human activities are the main cause of overproduction and emission of sulfur dioxide, including fossil fuel combustion, manufacturing of chemical products, and the burning of biomass (Kettle and Andreae, 2000;Halmer et al., 2002;Dentener et al., 2006;Lee et al., 2008) and China is the Fig. 3(B). The trip intensity and daily concentration of PM 10 from January 12 to March 27, 2019 and those in 2020 during the same period in Shenzhen, Guangzhou, and Foshan, respectively. world's largest coal producer and consumer (Kurokawa et al., 2013;Kato et al., 2016). It can be seen that SO 2 concentrations in Guangzhou and Foshan from January 12 to March 26 in 2020 were reduced, indicating that factory shutdowns caused by the Chinese New Year holiday and epidemic prevention actions led to significant reductions in SO 2 emissions. SO 2 concentrations in Shenzhen from January 12 to March 26 in 2020 increased, which may have been a result of the polluted northern air brought by the monsoon.

CO Concentration
Carbon monoxide (CO) is a colorless, odorless gas that forms when the carbon in fuels does not completely burn. About 60% of CO emissions nationwide come from vehicle exhaust, and up to 95% is found in cities. Other sources include fuel combustion in industrial processes and natural sources such as wildfires (Kyrkilis et al., 2007). The daily concentration of CO and trip intensity in January 12 to March 27, 2019 and those over the same period in 2020 are presented in Fig. 3(D), respectively. As shown in Fig. 3(D), between January 12 and March 27, 2019, in Shenzhen, Guangzhou and Foshan, the CO concentrations ranged from 0.40-0.96, 0.48-1.28, and 0.48-1.20 ppm, and averaged 0.58, 0.80, and 0.81 ppm, respectively. While those during 2020 ranged from 0.40-0.80, 0.48-0.96, and 0.40-0.96 ppm, and averaged 0.49, 0.62, and 0.57 ppm, respectively. The average values of CO in Shenzhen, Guangzhou, and Foshan in 2020 decreased by 15.6%, 21.8%, and 29.3% compared with those during the same period in 2019. Based on the combined data from the three cities, from January 12 to March 26 2020, the average CO decreased by 22.9% compared with that during the same period in 2019. When the trip intensity of Shenzhen and Guangzhou were less than 4.0 and 4.3, the CO concentrations corresponding to the same trip intensity were higher in 2020 than in 2019. Carbon monoxide (CO) in the atmosphere is mainly derived from the incomplete combustion of fossil fuels and biofuels. These combustion processes also emit nitrogen oxides and volatile organic compounds . During the epidemic control period, the CO emitted by civil boilers and power stations was significantly affected. Therefore, the same travel intensity corresponded to higher CO concentrations in 2020. In Foshan, the CO concentrations corresponding to the same trip intensity were higher in 2019 than in 2020. Foshan is an industrial city, and most factories did not return to normal production capacity during the epidemic control period, which means that the amount of CO derived from industrial production was significantly reduced. It can be seen that CO concentrations in the three cities from January 12 to March 26, 2020 were obviously reduced, indicating that the strict isolation measures had a significant impact on the generation and transformation of CO.

NO2 Concentration
Nitrogen dioxide (NO 2 ) mainly comes from social production activities such as chemical plants and motor vehicles (Cheng et al., 2018). NO 2 levels are often low and pose little immediate threat to human health. However, NO 2 is a cause for concern because it plays an important role in the formation of ozone, particle pollution, smog and acid rain (Kyrkilis et al., 2007). The daily concentration of NO 2 and trip intensity in January 12 to March 27, 2019 and those during the same period in 2020 are presented in Fig. 3(E), respectively. As shown in Fig. 3(E), between January 12 and March 27, 2019, in Shenzhen, Guangzhou, and Foshan, the NO 2 concentrations were between 4.90 and 35.5 ppb, 7.80 and 52.1 ppb, 4.90 and 44.8 ppb, and averaged 12.2, 23.6, and 21.6 ppb, respectively. Those during 2020 ranged from 4. 90-19.5, 4.90-31.7, and 4.90-32.1 ppb, and averaged of 9.80, 16.5, and 14.5 ppb, respectively. These values were, 20.2%, 30.1%, and 33% lower than those of during January 12 to March 27, 2019. Based on the data from the three cities, from January 12 to March 26 2020, the average NO 2 decreased by 29.1% compared with that during the same period in 2019. When the trip intensity of Shenzhen, Guangzhou, and Foshan was less than 3.7, 4.0, and 1.6, respectively, the O 3 concentrations corresponding to the same trip intensity were higher in 2020 than in 2019. Nitrogen dioxide (NO 2 ) mainly Fig. 3(E). The trip intensity and daily concentrations of NO 2 from January 12 to March 27 2019 and those of 2020 during the same period in Shenzhen, Guangzhou and Foshan respectively. comes from social production activities such as chemical plants and motor vehicles (Cheng et al., 2018). When the concentration of NO 2 from the automobile exhaust was reduced due to the reduction in trip intensity, the NO 2 sources from the combustion of fossil fuels, such as power plants continued normal operations and produced atmospheric pollutants. NO x from vehicle emissions has a significant impact on the concentration of NO 2 in the urban atmospheric environment (Cheng et al., 2018). It can be seen that NO 2 concentrations in the three cities from January 12 to March 26, 2020 were obviously reduced, indicating that a series of measures for epidemic prevention reduced NO 2 industrial and traffic emissions.

O3 Concentration
Ozone is a gas composed of three atoms of oxygen. Ozone occurs both in the Earth's upper atmosphere and at the ground level (Kyrkilis et al., 2007). O 3 is generated by photochemical reactions and can increase susceptibility to respiratory infections (Wu et al., 2015). Nitrogen dioxide (NO 2 ) as the precursor of ozone, and it participates in the formation of generate nitrate (NO 3 -) (Bowman et al., 1994). The daily concentrations of O 3 and trip intensity from January 12 to March 27, 2019 and those during the same period in 2020 are presented in Fig. 3(F), respectively.
As shown in Fig. 3(F), between January 12 and March 27, 2019, in Shenzhen, Guangzhou, and Foshan, the O 3 concentrations ranged from 14.5-60.2, 5.10-75.1, and 3.70-73.3 ppb, and averaged 36.2, 32.0, and 29.1 ppb, respectively. Those in 2020 during the same period ranged from 21.5-60.2, 13.1-70.0, and 9.80-67.7 ppb, and averaged 36.1, 35.6, and 33.7 ppb, respectively. The average values of O 3 in Guangzhou and Foshan from January 12 to March 26, 2020 increased by 11.4% and 15.8%, respectively, compared to those in 2019, while Shenzhen remained unchanged. Based on the combined data from the three cities, from January 12 to March 26, 2020, the average O 3 increased by 8.41% compared with that during the same period in 2019. However, the O 3 concentrations corresponding to the same trip intensity were higher in 2020 than in 2019, which was due to the low concentration of nitrogen dioxide making the generated ozone cannot be effectively converted .
Ozone is a secondary pollutant controlled by meteorological conditions and precursor pollutants. The precursors are generally NO x , VOCs, CO, and CH 4 (Vingarzan 2004;Kleinman, 2005). Because there are large number of anthropogenic sources in urban and industrial areas, the concentrations of NO x and VOCs are high, and they are important precursors for controlling ozone concentration (Lal et al., 2000). It is well known that the reciprocal transformation of O 3 , NO and NO 2 in atmospheric conditions is generally dominated by the following reactions (Clapp et al., 2001): NO 2 + hν(+O 2 ) → NO + O 3 Although there were many precursors of O 3 in the atmosphere in 2019, the meteorological conditions were not conducive to O 3 generation, and the concentration of O 3 in the atmosphere was at low levels in 2019, so the concentration of O 3 in 2019 was lower than that in 2020.

Distribution of the Six AQI Classes
In order to study the distribution of AQI, this study carried out statistical analysis and comparison between January 12 and March 27, 2019 and those in 2020 during the same period, respectively. The distribution of the six AQI Classes in Shenzhen, Guangzhou, and Foshan were shown and compared in Fig. 4.
From January 12 to March 27, in Shenzhen, the daily AQI averaged 45.9 in 2019 and 42.2 in 2020. In Guangzhou, the daily AQI averaged 63.2 in 2019 and 51.8 in 2020. In Fig. 3(F). The trip intensity and daily concentrations of O 3 from January 12 to March 27, 2019 and those in 2020 during the same period in Shenzhen, Guangzhou and Foshan respectively. Foshan, the daily AQI averaged 60.1 in 2019 and 48.1 in 2020. The AQI level rankings of the three cities were as follows: Guangzhou > Foshan > Shenzhen, which showed that Shenzhen had the best air quality, while Guangzhou had the worst. Fig. 4 shows the distribution of the AQI classes for the three cities combination between January 12 and March 27 2019 and those in 2020. In the three cities, January 12 to March 27, 2019, the proportions of classes Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ, and Ⅵ were 45.2%, 50.4%, 4.4%, 0%, 0%, and 0%, respectively. During the same period of 2020, the proportions of classes Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ, and Ⅵ were 62.7%, 37.3%, 0%, 0%, 0%, and 0%. Compared to the same period in 2019, the proportions of class Ⅰ showed an obvious increase (from 45.2% to 62.7%), while the proportions of class Ⅲ decreased from 4.4% to 0%, between January 12 and March 26, 2020. This indicated that compared with the same period in 2019, the air quality in Shenzhen, Guangzhou, and Foshan between January 12 and March 26, 2020 improved very significantly. There may be two reasons for a decrease in the AQI in 2020. The first reason is that January 24, 2020  was the traditional Chinese New Year, so a large number of factories were closed for the holiday, which greatly reduced the industrial emission sources of air pollutants. Second, during the epidemic control period, local residents cancel unnecessary travel activities, which led to a substantial reduction in non-essential production activities and a marked decrease in travel frequency.

Five Top Daily AQI Values and Indicator Air Pollutants
The five highest daily AQI values from January 12 to March 27, 2017-2020 and related air pollutants in Shenzhen, Guangzhou, and Foshan are shown in Tables 1, 2 and 3, respectively.
The above results indicated that in Guangzhou, the average value of the 15 days with the highest AQI from January 12 to March 27, 2017-2019 (non-epidemic period) was 127.9, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 and O 3 were 86.7 µg m -3 , 121 µg m -3 , 5.60 ppb, 0.97 ppm, 47.4 ppb, and 55.1 ppb, respectively. The average AQI on the top five days with the highest AQI from January 12 to March 26, 2020 (epidemic prevention and control period) was about 83.4, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 36.8 µg m -3 , 61.0 µg m -3 , 2.52 ppb, 0.70 ppm, 25.4 ppb, and 50.0 ppb, which were 57.5%, 49.7%, 55%, 27.1%, 46.3% and 9.2%, lower than those in 2017-2019, respectively. In Table 2, it can be seen that the average AQI in 2020 was 83.4, which was 34.8% lower than that of in 2017-2019 (AQI = 127.9). Compared with the same period in 2017-2019, the air quality from January 12 to March 26, 2020 improved significantly.
Among the 60 days with the highest AQIs in the four year period under observation (2017)(2018)(2019)(2020), the proportion of the number of days with the highest AQIs were 48.3%, 30% and 21.7% in January, February and March, respectively. It shows that the higher AQIs were mainly in January when the temperatures were lower, which was consistent with the conclusion drawn earlier that the frequency of PM 2.5 as indicatory air pollutant has increased in January due to the low temperatures, and the frequency of O 3 as indicatory air pollutant has increased in February and March due to the a warm temperatures. In winter, a lower ground temperature will hinder the dispersion of air pollutants and lead to an increase of PM 2.5 concentration in the atmosphere in cold season. Lee et al., 2018;Wang et al., 2018a). Under sufficient solar radiation intensity, NO 2 acts as the precursor of photochemical reactions, and first decomposes into NO and O (3P): The effect of radiation on atmospheric photochemical reactions deserves further study. When the ambient temperature is low, the three cities should pay attention to the impact of atmospheric particulates on human health. Fig. 5 show the distribution of the monthly average nearsurface streamlines in Guangdong Province (including Guangzhou, Foshan, and Shenzhen City) from January to March in 2019 and 2020. Fig. 6 shows that the monthly average wind speed in this region from January to February in 2019 was generally higher than the same period in 2020, but slightly lower in March. The monthly average wind speed in January 2019 was significantly higher than that in February and March in 2019 (about 1-1.5 m s -1 higher). During January to March of 2020, the monthly average wind speed in January was just slightly higher than that in February and March of 2020 (about 0.5 m s -1 higher). In terms of geographical distribution, Guangdong Province is a coastal province in southern China on the north shore of South China Sea, and the Pearl River flows through Guangdong Province from west to east and into the South China Sea in Central Guangdong.

Wind Streamlines and Wind Speeds
The Pearl River Delta Region (PRD, including Guangzhou, Foshan, and Shenzhen City) is a low-lying area surrounding the Pearl River estuary. Guangzhou City is located north of the center of the PRD; Foshan City is close to the west side of Guangzhou City, and Shenzhen is located at the southeastern corner of the PRD and close to the north side of Hong Kong. During January, because of the influence of the Siberia High, the northern corner of the PRD (about 2.0-3.0 m s -1 in 2019, 1.0-1.5 m s -1 in 2020) has relatively stronger wind speeds than the southern corner (below 1.5 m s -1 in 2019 and 2020). Until March, due to weakening of the Siberia High, the coastal region (southern corner of the PRD, below 1.5 m s -1 in 2019, 2.0 m s -1 in 2020) has relatively stronger wind speeds compared to the interior region (northern corner of the PRD, below 1.0 m s -1 in 2019, 1.5 m s -1 in 2020).
The distribution of the monthly average streamlines (Fig. 5) shows the places where the confluences of the streamlines generally have lower wind speed. In general, there are two main sources of airflow in Guangdong Province during from January to March. One of main sources of airflow is from inland. Due to the influence of the Siberian high pressure system, the prevailing inland airflow is a northerly (N) to northeasterly wind (NE), which is relatively strong in January, but gradually weakens with time. The other airflow is from the ocean. The prevailing oceanic airflow is an easterly (E) to east-northeasterly wind (ENE), which is weaker in January, but gradually increases with time. In the northern PRD (Guangzhou, Foshan City), the streamlines are mainly affected by the inland airflow in January, and the prevailing wind is a northerly wind (N) to north-northeasterly wind (NNE). From February to March, the streamlines in the northern PRD are affected by both a gradual weakening of the inland airflow and a gradually increases in the oceanic airflow. The confluence of the two airflows (inland and oceanic airflow) in the northern PRD, results in a lower wind speed in this region, making it easier to accumulate local air pollutants. During this period, Foshan is downwind of Guangzhou. In the southern PRD (Shenzhen City), from January to February, because of the confluence of the two airflows (inland and oceanic airflow), the wind speed in this region is lower, making it easier to accumulate local air pollutants. During March, the streamlines of the southern PRD are mainly affected by the oceanic airflow, and the prevailing wind is an easterly to northeasterly wind. At this time, Foshan and the south half of Guangzhou are the downwind of Shenzhen, so air pollution from Shenzhen to Guangzhou and Foshan along with the airflow.

CONCLUSIONS
1. In the combined the data from the three cities, from January 12 to March 27, in 2019, the trip intensity ranged between 0.94 and 4.82, and averaged 3.59, while those in 2020, the trip intensity ranged between 0.73 and 5.54, and averaged 2.57, which was 28.4% lower than that in 2019. This indicated that the strict prevention and control measures affected the trip plans of local residents. The trip intensity in these cities remained stable around its lowest point from February 4 to February 8, 2019, while in 2020, the lowest point was between January 25 and February 9, which was due to the implementation of strict epidemic control measures in Guangdong Province has led to a reduction in the frequency of unnecessary travel. 2. As to the three cities, from January 12 to March 27 of 2019 and 2020, the average concentrations of PM 2.5 , PM 10 , SO 2 , CO and NO 2 were 31.4 µg m -3 , 52.8 µg m -3 , 2.26 ppb, 0.73 ppm and 19.1 ppb in 2019, and 22.7 µg m -3 , 37.8 µg m -3 , 2.00 ppb, 0.56 ppm and 13.6 ppb in 2020, which were 27.7%, 28.5%, 11.8%, 22.9% and 29.1% lower than those in 2019, respectively. During the epidemic control period, restrictions on industrial production and mobile transportation resulted in reduction the average concentrations of PM 2.5 , PM 10 , SO 2 , CO and NO 2 . However, the average O 3 concentration (35.2 ppb) from January 12 to March 26, 2020 did not show a significant decrease, but rather increased by 8.4%. This is because