Levels, Chemical Compositions, and Sources of PM2.5 of Rural and Urban Area under the Impact of Wheat Harvest

Received: February 3, 2021
Revised: April 3, 2021
Accepted: April 7, 2021

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Li, Z., Yue, Z., Yang, D., Wang, L., Wang, X., Li, Z., Wang, Y., Chen, L., Guo, S., Yao, J., Lou, X., Xu, X., Wei, J. (2021). Levels, Chemical Compositions, and Sources of PM_2.5 of Rural and Urban Area under the Impact of Wheat Harvest. Aerosol Air Qual. Res. 21, 210026. https://doi.org/10.4209/aaqr.210026

HIGHLIGHTS

• Urban PM₅ was largely affected by wheat harvest (WH).
• Biomass burning (BB) markers increased after WH.
• Fugitive dust and BB have high and low contributions during WH.
• Coal was still an important contributor for urban and rural area.
• Higher vehicle exhaust shares occurred in rural area.

Keywords: Wheat harvest, PM2.5, Elements and ions, PMF, Beijing-Tianjin-Hebei

1 INTRODUCTION

North China Plain (NCP), one of two largest agricultural bases in eastern China, where the behaviors include winter wheat harvest in June and subsequent open burning of agricultural residues in field for maize planting are common. Open burning of wheat straw prevails in NCP due to low cost, convenience and insecticidal effect despite the straw burning prohibition in recent China (Wang et al., 2015; Li et al., 2020a, 2020b; Zhang et al., 2020). Large amounts of fugitive dust release from the disturbance of dust absorbed in wheat straw and surface soil by harvesters, wind, and cultivators (Li et al., 2020a, 2020b). Qiu and Pattey (2008) reported that the emission rate and factor for PM₁₀ during WH period were up to 0.58 ± 0.12 g m^–2 and 0.74 ± 0.12 kg ha^–1, respectively. Exposure to high level of agricultural dust can result in adverse health effects including acute and chronic bronchitis, mucous membrane irritation, allergic asthma and hypersensitivity, etc. (Ou et al., 2020; Rao et al., 2020). Also the subsequent straw burning for maize planting can result in the releasing of drastic amounts of pollutants such as CO₂, CO, NH₃, non-methane hydrocarbon, elemental carbon, organic carbon, and particulate matter, which influence air quality, air visibility, human health, and the climate (Chen et al., 2018; Zhang et al., 2020). The carbonaceous aerosols related to biomass burning contribute approximately 42% and 74% of black carbon (BC) and organic carbon (OC) globally, in which open burning of biomass accounts for 95% and 88%, respectively (Yao et al., 2017). More than 50% emissions of PM_2.5, OC, EC, K, K⁺, and Cl^– in eastern China was associated with wheat straw burning (Li et al., 2014).

As a large agricultural country, China contributed more than 24% of global carbonaceous aerosols, especially from field burning of biomass. The particulate matter (PM) emissions related to agricultural operations has constituted an emerging issue about air quality, especially in the areas between rural and urban environments (Qiu and Pattey, 2008; Wang et al., 2015). Biomass burning has a significantly adverse effect on regional air quality, global climate change through complex feedbacks with radiation and clouds, and public health ascribed to exposure to the derived gaseous and aerosol particles (Wang et al., 2015; Li et al., 2020a, 2020b; Liu et al., 2020).

Winter wheat mainly planted in North China Plain especially in Beijing-Tianjin-Hebei (BTH) region. The straw burning for maize plants in June after wheat harvest was common behavior because of its low cost, convenience, and insecticidal effect (Wang et al., 2015; Li et al., 2020b). Li et al. (2020a, 2020b) reported the transport mode of air pollutants as “rural to urban” for particulate matters during wheat harvest and organic pollutants after wheat harvest due to the open burning of wheat straw.

To our knowledge, the current studies mainly focused on the emission factors of PM during wheat harvest or the impact of subsequent biomass burning on local and urban areas (Qiu and Pattey, 2008; Yao et al., 2017; Chen et al., 2018).

 
2 METHODOLOGY

 
2.1 Description of PM_2.5 Sampling Sites and Campaign

The sampling locations had been shown in Li et al. (2021). Two collected sampling sites locate within Baoding Prefecture, which are an urban area (UA; 115°31ʹE, 38°53ʹN) in Baoding city and an adjacent rural site (ARS; 115°8ʹE, 38°40ʹN) in Wangdu County. Both UA and ARS locate in the central area of Beijing-Tianjin-Shijiazhuang triangle area. ARS within Wangdu County locates 58 km south of Baoding city. Wangdu County (115°01ʹ–115°18ʹE, 38°30ʹ–38°48ʹN), covers an area of 374 km² and has a permanent population of 275,000 (including 229,000 agricultural population), which is a big agricultural county with the wheat planting area of 18,095 hectares.

Two medium-volume air samplers (TH-150C) setting at 100 L min^–1 were used to consecutively collect PM_2.5 samples from June 18 to July 7 for both UA and ARS, and each sampling process lasted for 23 h. The quartz filter (QF) with 90 mm diameter was used to the analysis of 9 inorganic ions including Ca²⁺, K⁺, Mg²⁺, Na⁺, NO₃^–, SO₄^2–, Cl^–, F^–, and NH₄⁺. Teflon filter (TEF) was adopted to calculate the PM_2.5 mass by subtracting its pre-weight from post-weight and analyze 39 inorganic elements (IEs): Li, Be, Na, P, K, Sc, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Y, Mo, Cd, Sn, Sb, Cs, La, Ce, Sm, W, Tl, Pb, Bi, Th, Si, Al, Ca, Mg, Fe, Ti, Ba, Sr, and Zr. A series of field blank experiments were conducted to correct subsequent analysis deviation and ensure its accuracy. Both QFs and TEFs were purified prior to PM_2.5 sampling. The heating temperature was set as 450°C and 60°C for QFs and TEFs before PM_2.5 collections. A room with constant temperature of 20°C and relative humidity of 50% was used to store filters before and after sampling aiming to the obtainment of PM_2.5 mass.

 
2.2 Analysis of PM_2.5-associating Inorganic Elements

The detailed detection procedure documented in Li et al. (2019) was adopted in this study. An inductively coupled plasma mass spectrometry (ICP-MS) system (7500A Series; Agilent, USA) was used to detect 30 IEs including Li, Be, Na, P, K, Sc, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Y, Mo, Cd, Sn, Sb, Cs, La, Ce, Sm, W, Tl, Pb, Bi, Ti and Th. The rest 7 IEs involving Si, Al, Ca, Mg, Fe, Ba, and Sr were analyzed by an inductively coupled plasma atomic emission spectroscopy (ICP-OES) system (5100 Series; Agilent, USA). For ICP-MS system, the half TEF filter was heated using the both aqua regia and hydrofluoric acid at 120°C and lasted for 2 h, and then ramped to 130°C for dryness. Then it was heated again with HCl acid and stored in a plastic comparison tube for analysis. For ICP-OES, the other half TEF filter was heated at 300°C and then raised to 550°C for completely ashing. Then it was heated together with added absolute ethanol and NaOH, and boiled with water later. Finally, 2 mL HCl was added and diluted to 10 mL for analysis. This study used the standard soil materials named as GBW07446–07457 purchased from Chinese Center for National Standard Material to ensure the accuracy of ICP-MS and ICP-OES. The method detection limits fluctuated from 0.10 ng of Be to 1800 ng of Fe. The relative standard deviation (RSD) values were in the range of 2.12–6.15%.

 
2.3 Analysis of PM_2.5-associating Water-soluble Ions

In regard to water-soluble ions (WSIs), half QF filter was firstly put into the centrifuge tube and underwent 20 min ultrasonic extraction with 10 mL ultra-pure deionized water for three times, and then filtered by a TEF membrane with pore size as 0.22 µm (Whatman; Middlesex, UK). Finally, the extract was transferred into a plastic bottle and stored at 4°C before analysis.

ICS-1000 system (Thermo Scientific, USA) was used to detect SO₄^2–, NO₃^–, Cl^–, and F^–, and their MDLs were in the range of 0.005–0.012 µg mL^–1. K⁺, Na⁺, Ca²⁺, and Mg²⁺ were analyzed using ICP-OES (725 Series; Agilent, USA) with the MDLs fluctuating from 0.002 to 0.025 µg mL^–1. NH₄⁺ was measured using the ultraviolet-visible (UV–Vis) spectrophotometer (T6; Beijing General Instruments Co., Ltd) and its MDL was 0.01 µg mL^–1. Standard solutions corresponding to these 9 ions were obtained from the Research Center of China National Standard Reference Materials.

 
2.4 Source Apportionment of PM_2.5 by Positive Matrix Factorization

Positive Matrix Factorization (PMF) is a convenient multivariate factor analysis method based on non-negative constraints to get more practical factors (Yao et al., 2016). PMF version 5.0 available at the U.S. EPA website (https://www.epa.gov/air-research/positive-matrix-factorization-model-environmental-data-analyses) was applied to evaluate PM_2.5 potential source contributions. The input data profiles include concentration and uncertainty values for 39 inorganic elements and 9 water-soluble ions. The abnormal concentration values were excluded to prevent error and uncertainty was calculated as Eqs. (1) and (2):

The half of MDL was utilized to represent the corresponding mass content below MDL for chemicals. The geometric mean of the observed values was used as substitutes for missing values and four times the geometric mean represented their uncertainties (Zhang et al., 2015; Yao et al., 2016; Zong et al., 2016). A total of 20 runs were conducted for each chemical component. The lowest Q_robust value was 3088.20, and the ratio of Q_robust/Q_true was 0.91.

A total of six emission sources including biomass burning (BB), coal combustion (CC), industrial sources (ISs), vehicle exhaust (VE), fugitive dust (FD), and secondary inorganic aerosol (SIA) were discriminated by PMF.

The BB was characterized by high loadings of Cl^–, K, and K⁺. He et al. (2017) and Feng et al. (2018) reported that K⁺ and NH₄⁺ were commonly used markers for BB, and Cl^– was also used to identify the activities of wood burning and agricultural residues. The factor with high contributions of Cu, Zn, and Pb was recognized as VE. Cu, Zn, and Pb were widely used as markers for VE (Xue et al., 2019). The motor brake wears were often associated with Cu and Pb, while Zn was related to tire wears (Liu et al., 2015). Pb could no longer be act as a single VE marker because of its sharply decreased concentrations owning to the prohibition of Pb containing gasoline (Li et al., 2020b). Coal combustion profile is expressed by high loadings of As, Sb, and Cd, and medium concentrations of Pb and Tl (Li et al., 2016). As acting as a unique tracer of CC is frequently documented (Tao et al., 2014; Li et al., 2016; Yao et al., 2016). Pb, Sb, and Tl often released from coal combustion despite they are typical crustal elements (Zhang et al., 2014; Zong et al., 2016). The factor possessing high levels of water-soluble ions including NH₄⁺, NO₃^–, and SO₄^2– is always associated with SIA (Yao et al., 2016). Suspended dust was common in cities and rural areas due to the dust from unpaved roads, construction sites, and bare soil around urban area heaped by wind. Fugitive dust in this study features by high contents of Al, Mg, Si, Ca, Ti, Fe, and Ba. Previous studies had used Al, Mg, Si, Ca, Fe, and Ti as markers for road dust and crustal soil (Harrision et al., 2012; Vossler et al., 2016). In UA, the ISs are marked by high contributions of a host metal elements including K, Cr, Mn, Cu, Zn, Rb, Mo, Cd, Co, Tl, Sn, Pb, and W. The elements Mn and Cr can be considered as the steel-associating metals (Tao et al., 2014; Li et al., 2016), while Zn and K are often related to iron and steel processing (Li et al., 2021). Cd and Co are also associated with steel processing or other metallic industrial processes (Mooibroek et al., 2011). Tl, Sn, Co and Cu were frequently used in electroplating industry, and Al and Cr were mainly used in small electroplating workshops (Lestari and Mauliadi, 2009). Fe, Mn, Zn, Cd, and Ni were also used as indicators of alloy industry. Li and Pb were often associated with battery manufacturing (Yu et al., 2011). A large number of electroplating and battery enterprises in Baoding further proved this inference. However, the ISs in rural area of Wangdu County differed from those of UA, which possessed lower contributions of Li and Pb due to the lack of local battery factories.

 
2.5 Geo-accumulation Indexes of Inorganic Elements (IEs)

A parameter, the geo-accumulation index (I_geo), was widely used to evaluate the pollution degrees of IEs (Li et al., 2017, 2019), which could be obtained by Eq. (3), where C_n are the contents of elements in PM_2.5 and B_n are their soil background standard values in Hebei Province:

The pollution levels of individual elements including uncontaminated, uncontaminated to moderately contaminated, moderately contaminated, moderately to heavily contaminated, heavily contaminated, heavily to extremely contaminated, and extremely contaminated were identified according to the I_geo values of ≤ 0, 0–1, 1–2, 2–3, 3–4, 4–5, and > 5, respectively.

 
3 RESULT AND DISCUSSION

 
3.1 Levels of PM_2.5, Water-soluble Ions, and Elements of UA and ARS

Fig. 1 showed the mass and volume concentrations for inorganic elements and water-soluble ions during DWH and AWH for ARS and UA. PM_2.5 concentrations of UA fluctuated from 23.4 µg m^–3 of June 22 to 200 µg m^–3 of June 18 with a mean value as 87.7 µg m^–3, and which ranged from 23.3 µg m^–3 of July 1 to 155 µg m^–3 of June 19 averaging at 66.3 µg m^–3. It should be noted that higher PM_2.5 levels appeared in ARS instead of UA indicating that the strong impact of wheat harvest activities on PM_2.5 level (Qiu and Pattey, 2008; Zhang et al., 2020; Li et al., 2021). Generally, PM_2.5 peaks appeared earlier in ARS than corresponding peaks of UA combining with air mass origin analysis of UA indicated that pollutant migration from rural to urban. Both UA and ARS possessed much higher PM_2.5 concentrations in DWH compared with AWH, which implied that the influence of running of wheat harvesters and wheat transportation (Li et al., 2020a, 2020b, 2021). The average PM_2.5 concentrations in DWH for ARS and UA were 156 and 137 µg m^–3, which were well above the threshold value of 75 µg m^–3 in the Chinese National Ambient Air Quality Standard Grade II (GB3095-2012). The PM_2.5 levels (reported in µg m^–3) in ARS and UA were much higher than 116 of Handan in 2014 (Meng et al., 2016), 108 of Hangzhou during 2014 to 2015 (Liu et al., 2015), 108 and 86.8 of Beijing in 2000–2009 and 2010–2016 (Xu et al., 2017; Zhu et al., 2018), 121 of Lanzhou in 2012 (Tan et al., 2017), 91.7 of Shijiazhuang in 2016 (Lang et al., 2018), 100.6 of Xinxiang in 2015 (Feng et al., 2018), 38.9 of Ningbo in 2014–2015 (Zhang et al., 2018), and 62.5 of a background site of North China Plain in 2011 (Yao et al., 2016). High PM_2.5 concentrations caused by wheat harvest behaviors should be paid more attentions under the background of present downward trend of PM_2.5 levels on a national scale (Li et al., 2020a, 2020b).

Fig. 1. (a) Time series of PM_2.5 mass concentrations and (b) average PM_2.5 levels during DWH and AWH for ARS and UA.

 
3.2 Time Series of Dust Levels and Trace Element Oxides

Concerning the mass contents, the sum of 39 elements and 9 ions ranged from 134 mg g^–1 of June 30 to 963 mg g^–1 of July 3 and averaged at 444 mg g^–1 for ARS. They varied from 224 mg g^–1 of July 5 to 951 mg g^–1 of June 23 with a mean value of 450 mg g^–1 for UA.

Si, Ca, Al, Fe, and Ti were major crustal elements and could be used to calculate the soil dust (SD) mass with the expression as [SD] = 1.89 × [Al] + 2.14 × [Si] +1.4 × [Ca]+1.43 × [Fe] + 1.66 × [Mg] + 1.67 × [Ti] (Su et al., 2017). Among crustal elements, the elements other than Si, Al, Fe, K, Ca, Mg, Na, and Ti were defined as trace elements. Trace element oxides (TEOs) were evaluated by TEO = 1.3 × [0.5 × (Sr + Ba + Mn + Co + Rb + Ni + V) + 1.0 × (Cu + Zn + Mo + Cd + Sn + Sb + Tl + Pb + As + Se + Ge + Cs + Ga)] (Zhang et al., 2013). Due to the lack of Se, Ge, Cs, and Ga, the TEO was underestimated slightly.

Figs. 2 and 3 showed the time series of soil dust mass shares and TEO contents in PM_2.5, and the corresponding values during DWH, AWH, and WSP. Dust shares declined significantly from 9.40% to 6.75%, and from 21.3% to 7.98% for UA and ARS with the transition from DWH to AWH. It clearly evidenced that the wheat harvest behaviors caused the re-suspending of the dust absorption in wheat straws.

Fig. 2. Mass fractions of soil dust in PM_2.5 during sampling period for UA and ARS.

Fig. 3. Mass contents of trace element oxides during every sampling day and periods for UA and ARS.

The dust mass contributions ranged from 3.06% of June 18 to 12.9% of June 19 and averaging at 7.17% during whole sampling period (WSP) for UA, while higher corresponding values varied from 3.07% of June 22 to 27.1% of June 18 with a mean value as 9.99% for ARS. Higher dust mass shares attributing to ARS reflected the emissions of large amounts of dust associating with wheat harvest. UA possessed much higher dust fractions of 9.40% than the 4.6% of a background site in NCP (Yao et al., 2016) and 6.4% of Chengdu in summer (Tao et al., 2014), which indicated the impact of wheat harvest activities of the adjacent rural areas on air PM_2.5 compositions in urban areas. The average dust fraction 21.3% of ARS during DWH was also significantly higher than 13.2% of spring with frequent high concentrations of fugitive dust in a background site of NCP (Yao et al., 2016), which further proved the strong influence of wheat harvest behavior.

However, TEO contents exhibited the reverse variation trends: Higher values occurred during AWH instead of DWH. The TEO contents increased from 0.079% to 0.240% for UA, and from 0.254% to 0.280% along with the transition from DWH to AWH. Large amounts of trace elements emitted from wheat straw burning resulted in high TEO contents during AWH. Tao et al. (2014) and Zong et al. (2016) indicated that the biomass burning was an important source of trace elements. The TEO contents in UA and ARS were much lower than 7.2% of the adjacent Shijiazhuang in 2016, indicated the effectiveness of coal limitation and control of industrial emissions (Lang et al., 2016). At the same time, higher TEO levels belonged to ARS instead of UA further verified the validity of the reduction measures adopted in industrial sources. Li et al. (2020a, 2020b, 2021) implied large industrial emissions of TEO from industries.

 
3.3 I_geo Values for the Inorganic Elements

PM_2.5-bounded heavy metals (HMs) and As could threaten the human health and have been widely concerned (Mario et al., 2018; Li et al., 2019). Table 1 showed the I_geo values for 39 inorganic elements for UA and ARS. The crustal elements such as Al, Mg, Ca, Fe, and Ti for both UA and ARS possessed high I_geo values, which varied from 3.31 to 6.26 for ARS, and from 1.62 to 5.38 for UA (Table 1). Higher I_geo values of crustal elements in ARS reflected the impact of WH-related activities. Also higher Cu and Zn I_geo values in ARS suggested the influence of running of agricultural machineries. Considering HMs, Cd, Sb, Sn, Mo, and Pb possessed high I_geo values reflecting the impact of industrial production processes and coal combustion.

 
3.4 Levels of the Water-soluble Ions

The sum of mass contents of 9 WSIs were in the range of 0.051 g g^–1 of June 30 to g g^–1 of June 21 averaging at 0.321 g g^–1 for UA. The corresponding values for ARS varied from 0.095 g g^–1 of June 30 to 0.871 g g^–1 of July 3 with a mean value as 0.394 g g^–1. SO₄^2–, NO₃^–, and NH₄⁺ (SNN) dominated in 9 WSIs for both UA and ARS; SNN contributed 92.0% and 90.3% to total mass of 9 WSIs for UA and ARS, respectively. No significant differences concerning the mass contents of SO₄^2–, NO₃^–, and NH₄⁺ existed between DWH and AWH for both UA and ARS, which reflected the emissions sources of them were similar between these two periods (Fig. S1). The mass ratios of NO₃^–/SO₄^2– during DWH and AWH were 0.993 and 0.832 for ARS, and 1.21 and 0.718 for UA, which indicated that coal combustion except for DWH of UA was still an important source despite the coal limitation policy in China (Zhai et al., 2019).

Acting as the markers of biomass burning, Cl^– and K⁺ increased along with the transition from DWH to AWH for both UA and ARS (Fig. 4). ARS possessed higher contents of these two ions than UA indicating the strong influence of local biomass burning during AWH in rural area. Concerning UA, Cl^– contents fluctuated from 3080 to 3510 µg g^–1 during DWH, and from 1390 to 29,100 µg g^–1 during AWH. For ARS, they ranged from 3870 to 6300 µg g^–1 for ARS during DWH, and from 1710 to 18,000 µg g^–1 during AWH. K⁺ contents varied from 2960 to 4210 µg g^–1 and 2040 to 35,900 µg g^–1 during DWH and AWH for UA, and from 3930 to 8050 µg g^–1 and 2310 to 21,600 µg g^–1 for ARS. However, the mass concentrations (reported in µg m^–3) of Cl^– and K⁺ showed the reverse trends: Higher values occurred during DWH for both UA and ARS attributing to high PM_2.5 mass concentrations during DWH. Cl^– and K⁺ mass concentrations for UA during DWH were 0.466 and 0.490 µg m^–3, which were almost two times 0.241 and 0.274 µg m^–3 for AWH. Higher corresponding values of Cl^– and K⁺ appeared in ARS, which were 0.755/0.800 and 0.629/0.493 µg m^–3 during DWH and AWH, respectively. The mass concentrations of Cl^– and K⁺ (reported in µg m^–3) for both UA and ARS were much lower than 3.69 and 1.18 for Shijiazhuang in autumn of 2016 (Lang et al., 2018), and 2.0 and 2.3 for Beijing in summer of 2012 (Zhang et al., 2016). Lower Cl^– and K⁺ contents than some cities implied the effectiveness of emission control of related sources such as biomass burning and coal combustion (Li et al., 2020a, 2020b, 2021; Zhai et al., 2019).

Fig. 4. Contents of Cl^– and K⁺ during each sampling day and period for UA and ARS.

SNN mass concentrations also exhibited the same trends with Cl^– and K⁺: DWH possessed higher values due to its higher PM_2.5 levels compared than AWH. DWH possessed higher SO₄^2–, NO₃^–, and NH₄⁺ mass concentrations (reported in µg m^–3) 21.7, 21.6, and 14.6 for ARS, and 15.0, 18.2 and 11.8 for UA. Lower corresponding values during AWH 9.11, 7.58, and 6.51, and 6.19, 4.44, and 4.32 appeared in ARS and UA, respectively. The SNN concentrations in UA and ARS were lower than 28.4, 34.6, and 13.1 of Beijing in summer of 2012 (Zhang et al., 2016).

 
3.5 Source Apportionment by PMF for UA and ARS

Fig. 5 showed the source fractions during DWH, AWH, and WSP. It should be noted that BB shares in ARS increased from 5.27% to 25.5% along with the DWH to AWH, and the corresponding shares enhanced from 1.68% to 14.4% for UA. The similar increasing trends between UA and ARS, and higher BB shares in ARS proved the impact of open biomass burning for maize planting after wheat harvest in rural areas (Li et al., 2020a, 2020b, 2021). Higher contribution of FD was attributed to ARS, which was 22.81% during whole sampling period. Under the wheat harvest activities, the adjacent UA also possessed considerable FD contribution of 22.0% during WSP. Higher FD share 26.8% occurred during DWH for ARS, while it was 24.5% during AWH for UA. More direct influence in rural area from WH-related activities such as running of wheat harvesters and wheat transportation was the explanation of higher FD shares during DWH for ARS. The surprised higher FD during AWH share 24.5% occurred in UA instead of ARS (22.1%) suggesting that urban areas possessed some unique FD sources such as extensive construction of buildings and road dust. At the same time, due to the highly intensive operation of large amounts of wheat harvesters and wheat transportation tractors, ARS possessed higher VE contributions than UA during DWH. However, VE shares during AWH in UA were higher than ARS, which might be associated with the decrease of operation intensity of harvesters. It was important to note that VE still occupying a high share during AWH for ARS, which might be related to the running of rotary cultivators for maize planting and vehicles for fertilizer transportation. Much higher IS contribution 30.9% occurred in UA during DWH and subsequently decreased to 19.9% due to the increasing shares of BB and FD. High IS shares should be concerned by local government, and strict emission standards and appropriate control measures were urgent to developed. The secondary inorganic aerosol was a key contributor for UA and ARS during both DWH and AWH. CC contributed 12.0% and 9.66% to the total PM_2.5 mass for UA and ARS indicating coal was still an important fuel used in events including urban industrial production and rural cooking (Li et al., 2020a, 2020b, 2021).

Fig. 5. The contributions of six sources to the PM_2.5 for ARS and UA.

The contributions of FD and BB in UA and ARS associating with wheat harvest were much higher than Shijiazhuang, background site of North China Plain, Chengdu, Nanjing, and Lanzhou (Table S1). The VE possessed higher shares in UA and ARS than abovementioned cities due to the operation of wheat harvesters and rotary cultivators.

 
4 CONCLUSIONS

To evaluate the effects of wheat harvesting in rural regions on the air quality in adjacent urban areas, we conducted a systematic PM_2.5 sampling campaign at both a rural and an urban site (ARS and UA, respectively) in the center of the Beijing-Tianjin-Hebei megalopolis during and after the harvest. We then analyzed the samples to evaluate 1) the PM_2.5 concentrations; 2) the chemical components, specifically, 39 IEs and 9 WSIs; and 3) the differences between the two sites in terms of source contributions.

• We observed higher PM₅ concentrations at ARS than UA, but both sites showed elevated levels during DWH (156 and 137 µg m^–3 at ARS and UA, respectively) compared to AWH (75.6 and 53.1 µg m^–3, respectively). In addition, the PM_2.5 peaks appeared later at UA than ARS, demonstrating the rural-to-urban migration of pollution. Both locations also displayed higher I_geo values for the crustal elements than the other elements, confirming that the dust emissions originated from the wheat harvest.
  
  
• Based on the IE content, dust composed 3% and 9.40% of the PM_2.5 mass at ARS and UA during DWH, respectively, but only 7.98% and 6.75% during AWH, further confirming the migration effect. By contrast, the mass ratios of the TEOs to the PM_2.5 increased from 0.254% to 0.280% and from 0.079% to 0.240% at ARS and UA, respectively, between DWH and AWH, because of straw burning emissions following the harvest.
  
  
• The levels of the biomass markers Cl^– and K⁺ rose significantly from DWH to AWH at both sites, which is attributable to the open burning of straw for maize planting. Specifically, the Cl^– concentration varied 3870–6300 µg g^–1 at ARS and 3080–3510 µg g^–1 at UA during DWH but 1710–18,000 µg g^–1 and 1390–29,100 µg g^–1 during AWH, and the K⁺ concentration varied 3930–8050 µg g^–1 at ARS and 2960–4210 µg g^–1 at UA during DWH but 2310–21,600 µg g^–1 and 2040–35,900 µg g^–1 during AWH.
  
  
• PMF revealed that the contribution of BB grew from 5.27% to 25.5% and from 1.68% to 14.4% at ARS and UA, respectively, between DWH and AWH. Furthermore, the share from FD peaked during DWH at ARS (81%) but during AWH at UA, mainly owing to the rural-to-urban transport time. Unexpectedly, the operation of agricultural machinery led to a higher proportion of VE emissions at ARS than UA. Lastly, CC accounted for 9.66% and 12.0% of the total PM_2.5 mass at ARS and UA over the entire sampling period, respectively, proving that coal remains a common fuel for rural household cooking and urban industrial production.

 
ACKNOWLEDGEMENTS

This study was supported by the Beijing Natural Science Foundation (8212034), the Natural Science Foundation of Hebei Province (B2020502007), and the Fundamental Research Funds for the Central Universities (2020MS125).