Identification of dominant weather patterns and circulation regulation mechanisms of winter PM₂.₅ pollution in Iran

Meteorological conditions and atmospheric circulation play a crucial role in the formation, transport, and accumulation of air pollution, and machine-learning approaches provide an effective tool for classifying circulation patterns. This study investigates the dominant weather patterns and circulation regulation mechanisms of winter PM₂.₅ pollution in Iran. We integrate 2017–2022 in situ and gridded PM₂.₅ data, 1960–2025 ERA5 reanalysis, monthly GFDL-ESM4 PM₂.₅ simulations for 1960–2025, and Self-Organizing Maps (SOM) classification to identify pollution events, circulation patterns, and associated large-scale drivers. Results showed that Iran's annual mean PM₂.₅ concentration decreased at an annual rate of 4.21%, whereas high values remained concentrated in the northwest and central urban agglomerations. Extreme pollution events occurred mainly in winter (42.38%) and remained frequent in core cities. Circulation patterns were identified and linked to two core pollution modes: transport-dominated and accumulation-dominated pollution. The Arctic Oscillation (AO) was recognized as the primary large-scale factor regulating both modes through the excitation of distinct Rossby wave trains. Long-term ERA5 and GFDL-ESM4 analyses over the past six decades showed that Iran's mid–upper-tropospheric circulation has shifted from low to high geopotential height dominance, accompanied by a change in the AO–PM₂.₅ association around 2000 and a transition in the pollution-related mode from transport-dominated to accumulation-dominated. The findings provide scientific support for improved control of Iran's air pollution and more accurate early warning of heavy pollution episodes.