Data assimilation (DA) integrates the latest observational data into initial background field, producing an optimal analysis field in continuous time and space within the realm of numerical model forecasting to improve forecasting. However, its computational efficiency remains a concern due to its relatively slow processing speed. Alternatively, deep learning (DL) methods train assimilation models using observational or forecast data as labels, but its effectiveness is often constrained by the accuracy of these labels. In this paper, we proposed a novel approach that combined three-dimensional variational assimilation (3D-Var) with DL method, which integrated the Hadamard attention mechanism and Transformer modules into the Unet framework (HT-Unet) to assimilate sulfur dioxide (SO2) and improve the forecast skill. This hybrid method not only significantly improved the forecast accuracy but also accelerated computational processes. Three forecast experiments were conducted with background filed, 3D-Var, and HT-Unet analysis fields to assess forecast improvements in SO2 concentration. The results showed that the correlation with HT-Unet analysis field was nearly 50% higher than the background field, which was comparable to those from 3D-Var analysis field. The RMSEs of the SO2 concentration in the HT-Unet and 3D-Var DA experiments were reduced by 2.59 µg/m³ (29.50%), and 2.60 µg/m³ (29.61%). Additionally, the HT-Unet method achieved computational efficiency far surpassing traditional methods, which was 34 times faster than traditional approaches for assimilating a single analysis. The new method demonstrated the potential to replace the traditional 3D-Var method, overcoming the high-cost limitations of conventional variational assimilation and significantly enhancing air pollution forecasting accuracy.

Rapid increasing industries and city expansions have caused severe air pollution in the Guanzhong Basin (GZB), China in recent decades. Observations reveal that, although implementation of strict mitigation measures since 2014 has considerably reduced particulate matter (PM) pollution, the ozone (O) pollution during the warm season has continuously deteriorated in the basin. Simulations in May and August 2018 have been conducted using the WRF-Chem model to examine spatial and seasonal variations of the O formation regimes as well as source attributions in the GZB. The model generally performs well in simulating meteorological parameters, O, NO, and fine PM against measurements. The identified O formation regimes in cities of the GZB are all VOCs-sensitive in May and become more NO-sensitive in August. Sensitivity studies have shown that the power plants source generally suppresses the Oformation considerably in May and enhances it slightly in August due to its high NO and low VOCs emissions. The residential, transportation and industry sources increase the O concentration in May and August. Moreover, the transportation and industry sources play an increasingly important role in August but opposite for the residential source. The variation of O formation regimes and source attributions from May to August is caused by intensification of solar radiation, which not only promotes photochemical processes, also increases temperature and further enhances biogenic emissions and vertical exchange in the planetary boundary layer. The present study can provide guidelines to devise the effective O abatement strategies suitable for local situations.

In this study, we investigated the chemical composition and hygroscopicity of water-soluble fraction in PM2.5 collected from a rural site of Guanzhong Basin, a highly polluted region in northwest China. Hygroscopic growth factors, g(RH), of water-soluble matter(WSM) were measured by hygroscopic tandem differential mobility analyzer(H-TDMA) with an initial dry particle diameter of 100 nm. The g(90)WSM and κWSM was in the range of 1.08~1.49(1.35{plus minus}0.10) and 0.04~0.29(0.19{plus minus}0.06) in summer, 1.24~1.45(1.36{plus minus}0.07) and 0.12~0.26(0.20{plus minus}0.04) in winter, respectively. We found that increased nitrate concentration at night in summer suppressed 60-70% of the deliquescent point, and increased g(RH) at elevated relative humidity, compared to daytime. Secondary inorganic ions were the main components in heavy haze day, and greatly contributed to the hygroscopicity of particles. In contrast, more potassium compound and WSOM existed during Chinese Spring Festival event but exhibited no deliquescence point in the process of hygroscopic growth with the elevated RH. The g(90)WSOM and κWSOM, obtained using ZSR model, were in the range of 1.06~1.52(1.25{plus minus}0.14) and 0.024~0.32(0.13{plus minus}0.09) in summer, 1.06~1.58(1.38{plus minus}0.15) and 0.02~0.38(0.22{plus minus}0.10) in winter, respectively. The mean g(90)WSOM was in the range of that of biomass burning aerosols, and a good correlation (R=0.71) was found between g(90)WSOM and levoglucosan, confirming that the aerosol’s hygroscopicity were highly influenced by biomass burning in winter. Briefly, it is revealed that the aerosol in rural regions of Guanzhong Basin is mainly influenced by biomass burning based on the hygroscopicity in winter and summer.

Nitrate aerosols, formed via nitric acid (HNO) to balance inorganic cations in the particle phase, have constituted a major fraction of fine particulate matters (PM) during wintertime haze events in the North China Plain(NCP), with a progressively increasing contribution to PMmass. HNOis produced through homogeneous and heterogeneous pathways in the atmosphere, but the contribution of the two pathways to nitrate remains elusive. Simulations of a wintertime haze event in the NCP using a source-oriented WRF-Chem model reveal that the homogeneous and heterogeneous pathways contribute 48.4% and 51.6% of near-surface nitrate mass on average, respectively. The heterogeneous pathway dominates the nighttime HNOproduction in the planetary boundary layer, with an average contribution of 83%. Although NOis photolytically liable during daytime, the heterogeneous NOhydrolysis still contributes 10% of HNO. Our study highlights the significantly important role of NOheterogeneous hydrolysis in the nitrate formation during wintertime haze days.

Stringent mitigation measures have reduced wintertime PM2.5 concentrations by 42.2% from 2013 to 2018 in the BTH. The observed nitrate aerosols have not exhibited a significant decreasing trend and constituted a major fraction (about 20%) of the total PM2.5, although the surface-measured NO2 level has decreased by over 20%. It still remains elusive about contributions of nitrogen oxides (NOx) emissions mitigation to the nitrate and PM2.5 level. The WRF-Chem model simulations of a persistent haze episode in January 2019 in the BTH reveal that NOx emissions mitigation does not help lower wintertime nitrate and PM2.5 concentrations under current conditions in the BTH, because the substantial O3 increase induced by NOx mitigation offsets the HNO3 loss and enhances sulfate and secondary organic aerosols formation. Our results are further consolidated by occurrence of the severe PM pollution in the BTH during the COVID-19 outbreak with a significant reduction of NO2.