Figure 1 (a) Map showing the distribution of average surface SO2 mass concentrations in July in eastern China from OMI satellite reproduced dataset . (b) Measurement overview. Map showing the surface measurement sites (black circles with white numbers) and flight route (black dotted lines). The numbers from 1 to 9 correspond to the measurement sites in Heze, Jining, Liaocheng, Jinan, Zibo, Binzhou and 3 sites in Beijing. The spatial distributions of sea level pressure and 10-m wind field at 16:00 LT on July 28, 2018 are also shown corresponding to the areas in the black box marked in (a). The surface position of the warm front is marked with the symbol of a red line with semicircles. The meteorology data were downloaded from the ERA5 reanalysis website.
We conducted aircraft measurements of trace gases and aerosols from July to August 2018 in Jilin province of NE China. A Y-12 aircraft of the Jilin Weather Modification Office was used for the study. This twin-engine turboprop aircraft was similar to a Twin Otter. The aircraft campaign was conducted out of Baicheng city (122.84°E, 45.63°N), which is located 300 km from Changchun (the capital of Jilin, with a population of 7 million). The airport is located in the north (generally upwind direction) of the city. The sampling inlet was mounted at the bottom of the airframe with a forward-facing inlet connected to the measurement instruments. O3 and SO2 were measured with Thermo instruments (TEI 49i and 43i), and the aerosol scattering coefficients were measured with a commercial integrating nephelometer (TSI3563), which detects the light scattering coefficient (Bsp) at 450 nm, 500 nm, and 700 nm. All trace gas instruments were calibrated at the ground base and different altitudes on several selected flights. An Aerodyne time-of-flight Aerosol Chemical Speciation Monitor (TOF-ACSM) (Fröhlich et al., 2013; Sun et al., 2020) was deployed and fixed in the aircraft cabin for the chemical characterization of non-refractory PM2.5(NR-PM2.5). The detailed descriptions and data analysis of the instrument are in Text S1.
For multiple-surface measurements, the inorganic ions of PM2.5 were measured with a Monitor for Aerosols and Gases in ambient Air (MARGA, designed and manufactured by Applikon Analytical B.V., the Netherlands) (ten Brink et al., 2007) at 9 ground supersites in the NCP with three in the megacity Beijing and the other six in the Shandong Province, located from South to North in Heze, Jining, Liaocheng, Jinan, Zibo, and Binzhou, separately (Fig. 1b).
2.2 Designed aircraft experiments under the influence of mid-latitude cyclone
Previous studies have indicated the important role of mid-latitude cyclone in transporting pollutants from the NCP to NE China (Eckhardt et al., 2004; Stohl, 2001; Li et al., 2012; Dickerson et al., 2007; Oshima et al., 2013; Ding et al., 2009; Wu et al., 2018). In this study, we specially designed an aircraft experiment on July 30, 2018 under the influence of a cyclone. As shown in Fig. 1b, a deep mid-latitude cyclone was located over Mongolia (centered at 107°E, 45°N). The cyclone was characterized by a low-pressure system on the ground with a warm front extending from the center to the western edge of inner Mongolia. Wind vectors clearly show that warm and humid air from the NCP moved northwardly until encountered cold air mass from the north Mongolia and accordingly formed the convergence belt. Previous studies have identified the warm conveyor belt (WCB) from the distribution of 500 hPa specific humidity due to moisture advection associated with the lifting effect (Kiley and Fuelberg, 2006; Ding et al., 2009). The 500 hPa specific humidity field (Fig. S1) consistently shows a moist band over the front, revealing the strong vertical venting in this region. The distribution of specific humidity and wind filed indicate that a WCB-like circulation extended from the warm sector of the frontal system to NE China, where we conducted our aircraft study.
In addition to this specially designed experiment, we have also conducted 4 aircraft experiments respectively on August 2, August 3, August 8 and August 9. All the experiments were carried out during the daytime.
2.3 Positive matrix factorization (PMF)
Positive matrix factorization (PMF) (Paatero and Tapper, 1994) with the PMF2.exe algorithm was performed on the TOF-ACSM organics mass spectra to investigate various organic aerosol (OA) sources and processes. PMF analysis was performed with an Igor Pro-based PMF Evaluation Tool (Ulbrich et al., 2009), and the results were evaluated following the procedures detailed in Ulbrich et al. (2009) and Zhang et al. (2011). The total OA of all aircraft experiments was resolved into a hydrocarbon-like OA (HOA) factor and an oxygenated OA (OOA) factor. The detailed PMF results are described in Text S2.
2.4 The WRF-Chem model
The WRF-Chem model is an online-coupled chemical transport model that considers multiple physical and chemical processes, including emission and deposition of pollutants, advection, diffusion, gaseous chemical transformation, and aerosol chemistry (Grell et al., 2011). In this study, the WRF model (version 3.9.1) was employed to simulate weather conditions using the 1° x 1°NCEP (National Centers for Environmental Prediction) FNL Operational Global Analysis data (ds083.2). Meanwhile, the NCEP ADP Global Upper Air Observational Weather Data (ds351.0) was assimilated to improve the model’s meteorological reproduction.
The simulation is conducted from 1 July 2018 to 20 August 2018. The detailed configurations of the model are listed in Table S1. In this study, the model domain was centered at 35° N and 110° E with a 20 km × 20 km spatial resolution to cover the eastern China and its surrounding areas. On vertical distribution, 30 vertical layers are set from the surface to the top pressure of 50 hPa, and 10 of which are set below 1 km to better resolve the processes within boundary layer. Natural emissions such as biogenic, sea salt and dust are included online into the model runs for this study. For biogenic emissions, online biogenic emissions are calculated through MEGAN (Model of Emissions of Gases and Aerosols from Nature) module. It estimates the net emission rates of isoprene, monoterpene and other biogenic VOCs from terrestrial ecosystems into the above-canopy atmosphere. In addition to natural emissions, the anthropogenic emissions of CO, NOx, SO2, NH3, BC, OC, PM2.5. PM10 and VOCs are set based on the MEIC database (Multi-resolution Emission Inventory for China) (Huang et al., 2018).
To investigate the contributions of each individual physical and chemical process to variations of atmospheric nitrate and sulfate concentrations, we performed a diagnostic analysis in WRF-Chem modeling (Wang et al., 2020). The detailed diagnostic analysis and model evaluation are described in Text S3.
2.5 Lagrangian dispersion modeling
Lagrangian particulate dispersion modeling (LPDM) was used to study the transport pathways and trace the potential sources of air masses during the campaign using HYSPLIT (Stein et al., 2015), following the method developed by Ding et al. (2013). Briefly, for each hour during the study period, the model was run several days backwardly with 3,000 particles released every hour at a specific height over the site. The model calculated the particle position with mean wind and turbulent transport after being released from the receptor point. The residence time of particles at specific height columns was used to identify the footprint retroplume. The spatiotemporal distributions of these particles were used to identify the potential source regions and their relative contributions to the air masses at the measurement sites.

3. Results and discussions

3.1 Overall aircraft observation results
Fig. 2 shows the overall results of these 5 aircraft experiments conducted in the NEC. The overall NR-PM2.5 was dominated by secondary species, including sulfate, ammonium, nitrate and OOA since most of the flight time is outside the boundary layer, where the influence of surface primary emission is supposed to be small. Pollution episodes were captured by our flight on July 30 and August 3 with relatively higher mass concentrations of NR-PM2.5.
Sulfate, a secondary inorganic specie usually formed over a regional scale (Sun et al., 2011) contributed a significant fraction of NR-PM2.5 on July 30 (43%) and August 3 (54%). The sulfur oxidation ratios (defined as SOR = nSO42-/ (nSO42- + nSO2), where n represents molar concentration) were substantially higher on these two days, indicative of overall aged secondary aerosols. HOA, which represents primary organic aerosols (POA) usually associated with freshly traffic emissions (Zhang et al., 2011), contributed notably to NR-PM2.5 mass loadings during clean situations and its contribution were comparable on August 2 (13%), August 8 (15%) and August 9 (17%). While extremely lower contributions of HOA and relatively higher SOA to POA ratios were observed on pollution days, indicating the possible influence of regional transport rather than local sources. fm/z is defined as the fraction of the signal at the given m/z of the organic mass spectra derived from ACSM observation results. The organic aerosols of these two pollution cases (July 30 and Aug 3) both had higher values of f44, thus were more oxidized and aged, which was caused by extensive oxidation processes. These sulfate-dominated pollution episodes in the NEC are likely to be caused by regional transport due to extremely aged secondary aerosols while relatively lower values of f44and SOR together with notable fraction of HOA during clean situations indicate the importance of local sources.
Among these aircraft experiments, flight on July 30 was designed to investigate the role that mid-latitude cyclone plays in transporting air pollution and the aerosol chemistry alongside. Here we present the detailed analysis and in-depth understanding of this regional transport air pollution.