Severe cold air outbreak events over North America (CAO_NA) are often accompanied with a weakening of the stratospheric polar vortex. One of the main drivers causing this weakening is the intensified warm mass transport into the polar stratosphere (ST+). However, the robustness of the linkage between CAO_NA and ST+ events has yet to be fully established. Here, the empirical orthogonal function (EOF) method is applied to daily time series of mass transport into the polar stratosphere, using lead/lag days from −10 to 10 days around the peak dates of CAO_NA as the spatial domain and the number of individual CAO_NA events as the temporal domain. The results indicate that out of 227 CAO_NA events during the 41 winters (November–March) from 1979 to 2020, only 64 are not accom-panied with ST+ events. The remaining 72% (163) coupled with ST+ events can be categorized into three types: two types are characterized by ST+ events occurring a few days before (ST+_lead) or after (ST+_lag) the peak dates of CAO_NA events, while the third type corresponds to ST+ events occurring a week before (ST+_week) CAO_NA events. Both “ST+_week” and “ST+_lead” events are associated with a strengthening of wavenumber-1 waves, followed by a strengthening of wavenumber-2 waves. The “ST+_lead” and “ST+_week” events are featured by a “warm Arctic & cold midlatitudes” pattern in the Western Hemisphere with a reversed meridional dipole pattern in the Eastern Hemisphere. The “ST+_lag” events are purely driven by wavenumber-2 waves, corresponding to the co-occurring coldness over midlatitudes of both North America and Eurasia and a warmer Arctic. Further combining with the stratospheric polar vortex and circumpolar wind conditions in each type, a set of stratospheric weather regimes for CAO_NA events is identified.

Previous studies have reported a close relationship between the negative Arctic Oscillation (AO) and the PM_2.5 (particulate matter with a diameter of 2.5 μm or less) diffusion in North China in winter. Using the North China regional mean meridional wind at 850 hPa derived from the ERA5 (ECMWF Reanalysis version 5) reanalysis data in 1979–2022 as a useful substitute for station observed PM_2.5 concentration (since the latter is available only since 2014), our study detected strong/weak northerly events representing the abnormal PM_2.5 diffusion/accumulation events, and revisited the AO–PM_2.5 diffusion relationship in North China during 1979–2022. The results show that only when the AO was characterized by a 2-month continuously negative/positive phases and with twin peaks respectively before and after the diffusion/accumulation events, would there be higher occurrences of the abnormal PM_2.5 diffusion/accumulation. The second peak of negative AO acted to prolong the strong northerly winds by an average of 2 days. Further analysis reveals that the AO with twin peaks always has a footprint in the stratospheric northern annular mode (NAM) during the abnormal PM_2.5 events, and the coupling between the stratosphere and troposphere plays a critical role in the second peak of AO. Vertical propagation of baroclinically amplifying waves leads to changes in isentropic meridional mass fluxes in the stratosphere following the changes in the troposphere. The stronger/weaker poleward mass fluxes increase/decrease the polar mass in the stratosphere, which dominates the total column air mass changes and leads to the second peak of AO. Considering the subseasonal predictability of the stratospheric NAM based on existing evidence, particular attention should be paid to these AO-related abnormal PM_2.5 diffusion and accumulation events in North China because they might be more predictable at a longer lead time.