Solar and QBO influences on the timing of stratospheric sudden warmings

The interaction of the 11-yr solar cycle (SC) and the quasi-biennial oscillation (QBO) and their influence on the Northern Hemisphere (NH) polar vortex are studied using idealized model experiments and ECMWF Re-Analysis (ERA-40). In the model experiments, the sensitivity of the NH polar vortex to imposed easterlies at equatorial/subtropical latitudes over various height ranges is tested to explore the possible influence from zonal wind anomalies associated with the QBO and the 11-yr SC in those regions. The experiments show that the timing of the modeled stratospheric sudden warmings (SSWs) is sensitive to the imposed easterlies at the equator/subtropics. When easterlies are imposed in the equatorial or subtropical upper stratosphere, the onset of the SSWs is earlier. A mechanism is proposed in which zonal wind anomalies in the equatorial/subtropical upper stratosphere associated with the QBO and 11-yr SC either reinforce each other or cancel each other out. When they reinforce, as in S-min-QBO-east (S-min/E) and S-max-QBO-west (S-max/W), it is suggested that the resulting anomaly is large enough to influence the development of the Aleutian high and hence the time of onset of the SSWs. Although highly speculative, this mechanism may help to understand the puzzling observations that major warmings often occur in S-max/W years even though there is no strong waveguide provided by the QBO winds in the lower equatorial stratosphere. The ERA-40 data are used to investigate the QBO and solar signals and to determine whether the observations support the proposed mechanism. Composites of ERA-40 zonally averaged zonal winds based on the QBO (E/W), the SC (min/max), and both (S-min/E, S-min/W, S-max/E, S-max/W) are examined, with emphasis on the Northern Hemisphere winter vortex evolution. The major findings are that QBO/E years are more disturbed than QBO/W years, primarily during early winter. Sudden warmings in S-max years tend to occur later than in S-min years. Midwinter warmings are more likely during S-min/E and S-max/W years, although the latter result is only barely statistically significant at the 75% level. The data show some support for the proposed mechanism, but many more years are required before it can be fully tested.