Teleconnection between the Asian Polar Vortex and surface PM_2.5 in China

The study investigates how large-scale atmospheric circulation, particularly the Asian Polar Vortex (APV), influences the spatial and temporal variability of surface PM2.5 across China. PM2.5 affects human health, radiative forcing, and climate, and is a key air pollution indicator. China’s PM2.5 pollution exhibits strong regional variability, with persistent hotspots in the North China Plain (including Beijing–Tianjin–Hebei), Yangtze River Delta, Pearl River Delta, Guanzhong Plain, and Sichuan Basin. Traditional analyses focusing on site-level composition or local meteorology are limited for understanding nationwide, climate-scale drivers. Given the large spatial scales of synoptic systems, the authors analyze PM2.5 patterns over the entirety of China across 50 months (2015–2019), linking them to planetary-scale circulation via EOF decomposition and teleconnection with APV indices. The research seeks to quantify the extent to which APV variability explains PM2.5 variability and delineates how typical PM2.5 patterns correspond to distinct mid-tropospheric (500 hPa) and lower-tropospheric (850 hPa) circulation anomalies, thereby informing potential climate–air quality linkages.

Prior work shows meteorology can explain up to ~50% of daily PM2.5 variability in the U.S., indicating strong synoptic influences on air pollution. Climate change projections suggest increases in surface ozone by 1–10 ppb in polluted regions and changes in PM2.5 concentrations by 0.1–1 µg/m3 over coming decades. The East Asian winter monsoon is influenced by the Arctic Oscillation; Arctic amplification and sea-ice anomalies can modulate mid-latitude circulation persistence and extremes. The polar vortex, and specifically the Asian Polar Vortex (APV), affects temperature, humidity, precipitation, and thus aerosol formation, removal, and transport over China. Teleconnection frameworks have been widely used to link distant circulation anomalies to regional climate and dust activity. These insights underpin the hypothesis that APV variability is a key indirect climatic driver of PM2.5 distributions in China, beyond relatively stable spatial patterns of emissions.

Data: Raw hourly PM2.5 observations from the National Urban Air Quality Real-time Release Platform (MEPC) spanning January 2015–February 2019 were aggregated to daily 24-h geometric means. Meteorological fields (500-hPa and 850-hPa geopotential heights; H500 and H850) were obtained hourly from NCEP CFSv2 (0.5°×0.5°) and averaged to daily means. Monthly APV indices—Area Index (AIAPV) and Strength Index (SIAPV)—were obtained from the Climate Center of CMA. Preprocessing: For EOF analysis, daily PM2.5 at each site was detrended and deseasonalized by subtracting a 31-day moving average. For correlation with APV, monthly PM2.5 and APV indices were deseasonalized by subtracting nonoverlapping 3-month means and then normalized. EOF analysis: EOF decomposition was applied to the PM2.5 field to obtain dominant spatial modes (EOFs) and principal components (PCs). Eigenvalue significance was assessed; the leading four EOFs were retained. PCs were used to identify periods of typical PM2.5 patterns via extreme values (e.g., top/bottom 10%). Circulation composites: For periods when PCs exceeded thresholds (e.g., >1.280 or <−1.280), concurrent atmospheric circulation anomalies (500-hPa and 850-hPa geopotential heights) were composited to characterize circulation associated with each PM2.5 pattern. Teleconnection and variance attribution: Site-level correlations between deseasonalized monthly PM2.5 and AIAPV/SIAPV were computed. The percentage of PM2.5 variance explained by APV at each site was defined as 100% times the sum of squared correlation coefficients with AIAPV and SIAPV. Analysis focus: Emphasis was on large-scale circulation effects; local emission variations were not analyzed. Due to denser station coverage in eastern China, interpretations focus on central and eastern regions.

• Seasonal variability: Winter mean PM2.5 across China was 63 µg/m3 with 375 sites exceeding 75 µg/m3; summer mean was 28 µg/m3 with only 3 sites exceeding 75 µg/m3, mainly in North China. Large daily variability was observed, especially in heavily polluted North China.
• EOF patterns and variance: The leading four EOFs explained 50.5% of total variance. EOF-1 (28.4%) captured an overall pollution phase with widespread positive loadings and a high-value center in Central China. EOF-2 (9.7%) represented a north–south dipole (positive north of Qinling–Huaihe, negative to the south), with a positive center over Beijing–Tianjin–Hebei and negative over the Yangtze River Basin. EOF-3 (6.5%) showed an east–west phase (positive central China, negative northeast and eastern coastal areas) with a positive center in Guanzhong and negative in the Northeast Plain. EOF-4 (5.9%) reflected a north–center–south phase.
• Circulation linkage: Extreme values of PCs correspond to distinct 500-hPa and 850-hPa geopotential height anomalies. For example, when PC-1 > 1.280 (high national/ North China pollution), higher pressure over Siberia and lower pressure over the western Pacific and Tibetan Plateau were observed; opposite anomalies occurred when PC-1 < −1.280. When PC-2 > 1.280 (pollution concentrated in Beijing–Tianjin–Hebei and surroundings), dual high-pressure centers appeared over western China and Japan.
• APV correlations: Significant positive correlations existed between site PM2.5 and APV indices, stronger for SIAPV than AIAPV. There were 350 sites with correlation coefficient r > 0.5 for PM2.5 vs SIAPV and 126 sites with r > 0.5 for PM2.5 vs AIAPV, with higher correlations in North and Northeast China.
• Variance explained by APV: At 33 sites, >70% of PM2.5 variance was explained by APV; at 206 sites, >50%; and at 569 sites, >30%. The strongest APV control occurred in northern and northeastern regions, with more than 70% of variance explained around Beijing and its surroundings, particularly along the Bohai Sea and the Northeast Plain.

The analysis demonstrates that large-scale atmospheric circulation, encapsulated by the APV indices, exerts strong control on the spatial and temporal variability of PM2.5 across China, particularly in northern and northeastern regions. The dominant PM2.5 patterns derived via EOFs align with distinct mid- and lower-tropospheric circulation anomalies, indicating that synoptic to planetary-scale circulation states modulate pollution build-up, transport, and dispersion. The stronger association with SIAPV suggests that the strength of the APV more directly influences China's wintertime meteorology (e.g., temperature, pressure gradients, monsoonal flow) relevant to PM2.5 dynamics than its areal extent alone. These findings support the teleconnection between Arctic/mid-latitude circulation variability and Chinese air quality and imply that ongoing Arctic changes and APV variability can significantly impact PM2.5 pollution risk, especially over North China. The circulation composites tied to PC extremes provide mechanistic context linking observed pollution patterns to characteristic pressure anomalies and Rossby wave configurations.

This study establishes a robust teleconnection between the Asian Polar Vortex and surface PM2.5 in China. Four leading PM2.5 EOF modes capture 50.5% of variance, delineating national-scale and regional dipole/tripole patterns. Extreme occurrences of these patterns coincide with distinct 500-hPa and 850-hPa geopotential height anomalies. Deseasonalized monthly PM2.5 correlates significantly with APV indices (particularly SIAPV), and APV explains a substantial fraction of PM2.5 variance at many sites, exceeding 70% around Beijing and the Northeast Plain. These results highlight APV as a key climatic driver of PM2.5 variability in northern China and indicate that the findings can aid in anticipating the potential impacts of climate variability and change on air quality.

• The analysis focuses on large-scale circulation; local emission variations and changes are not explicitly analyzed or separated, despite their importance for absolute PM2.5 levels.
• Station density is higher in eastern China; results and interpretations primarily emphasize central and eastern regions.
• The study period spans 50 months (Jan 2015–Feb 2019), potentially limiting characterization of longer-term variability.
• The approach relies on statistical associations (EOFs, correlations, composites); it does not explicitly resolve chemical processes or quantify causality.
• Only PM2.5 mass concentrations were decomposed; no composition-resolved analysis was performed.