key: cord-0843697-4s4usyiq
authors: Zhang, Junke; Li, Huan; Chen, Luyao; Huang, Xiaojuan; Zhang, Wei; Zhao, Rui
title: Particle composition, sources and evolution during the COVID-19 lockdown period in Chengdu, southwest China: Insights from single particle aerosol mass spectrometer data
date: 2021-11-09
journal: Atmos Environ (1994)
DOI: 10.1016/j.atmosenv.2021.118844
sha: 89c8efd567b9435fb1c4d528871db1f9f98f5223
doc_id: 843697
cord_uid: 4s4usyiq

In order to investigate the effects of the Coronavirus Disease 2019 (COVID-19) lockdown on air quality in cities in southwest China, a single particle aerosol mass spectrometer (SPAMS) and other online equipments were used to measure the air pollution in Chengdu, one of the megacities in this area, before and during the lockdown period. It was found that the concentrations of fine particulate matter (PM(2.5)), nitric oxide (NO), nitrogen dioxide (NO(2)), sulfur dioxide (SO(2)) and carbon monoxide (CO) decreased by 38.6%, 77.5%, 47.0%, 35.1% and 14.1%, respectively, while the concentration of ozone (O(3)) increased by 57.5% from the time before to the time during lockdown. All particles collected during the study period could be divided into eight categories: biomass burning (BB), coal combustion (CC), vehicle emissions (VE), cooking emissions (CE), Dust, K-nitrate (K–NO(3)), K-sulfate (K–SO(4)) and K-sulfate-nitrate (K-SN) particles, and their contributions changed significantly after the beginning of lockdown. Compared to before lockdown, the contribution of VE particles experienced the largest reduction (by 14.9%), whereas the contributions of BB and CE particles increased by 7.0% and 7.3%, respectively, during the lockdown period. Regional transmission was critical for pollution formation before lockdown, whereas the pollution that occurred during the lockdown period was caused mainly by locally emitted particles (such as VE, CE and BB particles). Weighted potential source contribution function (WPSCF) analysis further verified and emphasized the difference in the contribution of regional transmission for pollution formation before and during lockdown. In addition, the potential source area and intensity of the particles emitted from different sources or formation mechanisms were quite different.

on air quality in cities in southwest China, a single particle aerosol mass spectrometer (SPAMS) and 43 other online equipments were used to measure the air pollution in Chengdu, one of the megacities in 44 this area, before and during the lockdown period. It was found that the concentrations of fine 45 particulate matter (PM2.5), nitric oxide (NO), nitrogen dioxide (NO2), sulfur dioxide (SO2) and carbon 46 monoxide (CO) decreased by 38.6%, 77.5%, 47.0%, 35.1% and 14.1%, respectively, while the 47 concentration of ozone (O3) increased by 57.5% from the time before to the time during lockdown. All 48 particles collected during the study period could be divided into eight categories: biomass burning (BB), 49 coal combustion (CC), vehicle emissions (VE), cooking emissions (CE), Dust, K-nitrate (K-NO3), 50 K-sulfate (K-SO4) and K-sulfate-nitrate (K-SN) particles, and their contributions changed significantly 51 after the beginning of lockdown. Compared to before lockdown, the contribution of VE particles 52 experienced the largest reduction (by 14.9%), whereas the contributions of BB and CE particles 53 increased by 7.0% and 7.3%, respectively, during the lockdown period. Regional transmission was 54 critical for pollution formation before lockdown, whereas the pollution that occurred during the 55 lockdown period was caused mainly by locally emitted particles (such as VE, CE and BB particles).

Weighted potential source contribution function (WPSCF) analysis further verified and emphasized the 57 difference in the contribution of regional transmission for pollution formation before and during 58 lockdown. In addition, the potential source area and intensity of the particles emitted from different 59 sources or formation mechanisms were quite different. 

Natick, MA, USA). The mass spectral data were grouped into different clusters using an adaptive 140 resonance theory neural network algorithm (ART-2a) with a vigilance factor of 0.6 and a learning rate 141 of 0.05 over 20 iterations (Hopke and Song, 1997; Song and Hopke, 1999) . Then, the first 158 clusters 142 generated by ART-2a (representing more than 96% of all the analyzed particles) were further manually 143 merged into eight particle types based on the similarity of chemical features and temporal variations.

Each particle type was named by its chemical composition, but the name does not represent all 145 chemical species in the single particle mass spectra. 

Meanwhile, the average mass concentrations of PM10 were also reduced, from 92.4 ± 31.8 μg m −3 to 195 52.8 ± 24.2 μg m −3 , and accordingly there was an obvious improvement in visibility from 6.9 ± 3.9 km 196 to 10.8 ± 4.4 km. The ratio of PM2.5 to PM10 has been used to evaluate the relative contributions of 197 fine/coarse particles to air pollution. In this study, compared to the period before lockdown (0.82), the 198 PM2.5/PM10 ratio increased to 0.88 during the lockdown period, which means that the contribution of 199 fine particles to total particles increased in Chengdu during the lockdown period. This can be attributed 

Over the whole study period, K-NO3, K-SO4 and K-SN particles constituted 21.7%, 3.5% and 296 23.8% of total analyzed particles, respectively. It is well known that nitrate and sulfate in atmospheric CE particles, all decreased. This suggests that the pollution episode that occurred before lockdown was 338 caused mainly by regional transmission. This is consistent with the formation of a typical pollution 339 process reported in a previous study in Chengdu (Zhang et al., 2021 

Compared with the whole study period, the trend in the change in secondary inorganic particles in 341 EP2 was completely opposite to that in EP1, and their total contribution decreased by 8.1%. In 342 particular, the contribution of K-NO3 particles decreased by 5.3%. However, the contributions of BB 343 particles from rural areas around Chengdu and CE particles from Chengdu city increased by 7.2% and 344 5.7%, respectively. Meanwhile, we found that, although the contribution of VE particles (16.1%) 345 during EP2 was lower than the average contribution over the whole study period (20.1%), it was 3.4% 346 higher than the average contribution during the lockdown period (12.7%). Therefore, VE particles also 347 made an important contribution to the formation of EP2. Thus, we can conclude that EP2 was caused In order to explore the differences in potential source areas of air pollutants in Chengdu before and 352 during COVID-19 lockdown in Chengdu, the WPSCF was used to simulate the source probability 353 distribution of PM2.5 and some single particle types in different periods. Here, in addition to the two 354 pollution episodes occurring before and during the lockdown periods (i.e., EP1 and EP2), one period 355 with a low PM2.5 mass concentration (January 31 to February 3, when the average concentration was 356 34.6 ± 4.4 μg m −3 , referred to as the "clean period") was also selected for a comparison of potential 357 source areas. Meanwhile, in addition to the PM2.5 mass concentration and total particle number 358 concentration ("Total"), one typical regional transmission particle type (i.e., K-SO4 particles) and one 359 typical local emitted particle type (i.e., VE particles) were selected for the WPSCF analysis, so as to 360 obtain more comprehensive results for potential source differences.

J o u r n a l P r e -p r o o f 

As shown in Fig. 5 , the potential source areas of various pollutants (PM2.5, "Total", K-SO4 and VE 365 particles) in the three periods (clean, EP1 and EP2) had obvious differences. During the clean period, 366 the potential sources of PM2.5 were distributed mainly in the area to the east of Chengdu, and the 367 corresponding WPSCF value was low (< 0.5), which means that the contribution of regional 368 transmission to PM2.5 in Chengdu was low during this period, and PM2.5 in Chengdu came mainly from 369 local emissions. There was good consistency in the potential source areas of "Total" and K-SO4 370 particles, which were concentrated mainly in the areas to the east and south of Chengdu. Although the 371 potential source area in the south was much smaller than that in the easterly direction, the WPSCF 372 value was much higher than that in the eastern area. Meanwhile, the potential source area in the eastern 373 area for "Total" and K-SO4 particles was larger than that in the eastern area for PM2.5. Although the 374 potential source areas of VE particles were also located in the areas to the east and south of Chengdu, 375 their potential source areas were significantly smaller than those of the other two types of particles 376 (especially in the easterly direction).

As mentioned above (section 3.3), EP1 was the pollution episode with the highest mean value of 378 PM2.5 and it had a long duration during the whole study period. Therefore, it can easily be found that 379 the potential source areas were more complex than in the other two periods, and the areas with high 380 WPSCF values (> 0.7) were widely distributed in the areas surrounding Chengdu. As shown in Fig. 5 

The potential source areas of "Total" particles during EP1 were distributed mainly in the area to the 385 northeast of Chengdu, and the areas with high WPSCF values were close to Chengdu. Similar to PM2.5, 386 the potential source areas of K-SO4 particles were distributed mainly in the areas to the northeast and 387 southeast of Chengdu and the junction area of Sichuan province and Chongqing, showing obvious 388 regional transmission characteristics, but the intensity of its WPSCF was far weaker than for PM2.5. The 389 potential source area of VE particles was the smallest of all pollutants, and its high WPSCF values 390 were concentrated mainly in the areas surrounding Chengdu, which is consistent with the fact that this 

During the COVID-19 Pandemic

HYSPLIT4 USER's GUIDE

Aerosol and 476 pollutant characteristics in Delhi during a winter research campaign

The role of primary emission 479 and transboundary transport in the air quality changes during and after the COVID-19 lockdown 480 in China

Spatiotemporal distribution of 482 satellite-retrieved ground-level PM2.5 and near real-time daily retrieval

Mineral 485 dust and NOx promote the conversion of SO2 to sulfate in heavy pollution days

Classification of single particles by neural networks based on the 487 computer-controlled scanning electron microscopy data

Enhanced secondary pollution offset 491 reduction of primary emissions during COVID-19 lockdown in China

Characterization of oxalic 493 acid-containing particles in summer and winter seasons in Chengdu

Water-soluble ions in PM2.5 during spring haze and dust periods in Chengdu

Variations, nitrate formation and potential source areas

Investigating the evolution of summertime secondary atmospheric pollutants in urban 500

Real time bipolar time-of-flight mass spectrometer for analyzing single aerosol particles

Air 506 quality changes during the COVID-19 lockdown over the Yangtze River Delta Region: An insight 507 into the impact of human activity pattern changes on air pollution variation

Characteristics and source 510 apportionment of PM2.5 during persistent extreme haze events in Chengdu

Quantitative assessment of atmospheric emissions of toxic heavy metals from anthropogenic 561 sources in China: historical trend, spatial distribution, uncertainties, and control policies

Changes in air quality during the lockdown in Barcelona (Spain) one month into the 565 SARS-CoV-2 epidemic

Suzhou in the Yangtze River Delta

Mixing state of individual carbonaceous 570 particles during a severe haze episode in

Long-range transport and regional sources of PM2.5 572 in Beijing based on long-term observations from

COVID-19 lockdown on ambient levels and sources of volatile organic compounds (VOCs

Severe air pollution events not avoided by 577 reduced anthropogenic activities during COVID-19 outbreak

Direct links 580 between hygroscopicity and mixing state of ambient aerosols: estimating particle hygroscopicity 581 from their single-particle mass spectra

Four-month changes in air quality during and after the COVID-19 lockdown in six megacities in 584

TrajStat: GIS-based software that uses various trajectory 586 statistical analysis methods to identify potential sources from long-term air pollution measurement 587 data

Mechanism for the formation of the January 2013 heavy haze pollution episode over central 590 and eastern China

Characteristics and formation mechanisms of autumn haze pollution in Chengdu based on high 595 time-resolved water-soluble ion analysis

Characterization of trace elements in PM2.5 aerosols in the vicinity of highways 597 in northeast New Jersey in the U

Refined source apportionment of coal combustion sources by using single particle mass 600 spectrometry

Single particle mass spectral signatures from vehicle exhaust particles and the source 605 apportionment of on-line PM2.5 by single particle aerosol mass spectrometry

Sources apportionment of PM2.5 in a background site in the North China Plain

Temporal variations in 611 the air quality index and the impact of the COVID-19 event on air quality in western China

Characterization, mixing state, and evolution of single particles in a 615 megacity of Sichuan Basin, southwest China

Insights into the 617 characteristics of aerosols using an integrated single particle-bulk chemical approach

Analysis of the characteristics of single atmospheric particles in Chengdu using 621 single particle mass spectrometry

Air pollution episodes during the COVID-19 623 outbreak in the Beijing-Tianjin-Hebei region of China: An insight into the transport pathways and 624 source distribution

Significant 627 changes in the chemical compositions and sources of PM2.5 in Wuhan since the city lockdown as 628 COVID-19

High-resolution sampling and analysis of ambient particulate 631 matter in the Pearl River Delta region of southern China: source apportionment and health risk 632 implications