Though not necessarily significant, the correlation was found negative between interannual anomalies of the AL and IL intensities for any winter month in any of the AGCM integrations. In many of the integrations the seasonality of the seesaw formation is not as distinct as in the observation for the recent decades. Still, five out of the eight hindcast integrations exhibit a clear tendency that the negative correlation becomes stronger towards the later part of the winter, in accordance with the observation.
A rather distinct interdecadal modulation was observed in the seasonality of the seesaw formation. The negative AL-IL correlation has been strongest in late February since the 1970s, and it was so in January during the 1950s, but the correlation was not significant in the 1960s. This observed modulation was not reproduced well in the hindcast integrations where no systematic interdecadal tendency was found in the AL-IL correlation. Interestingly, an clear interdecadal tendency in the AL-IL correlation was found in the control run. Reflecting this tendency, the leading EOF of the wintertime interannual variability over the Northern Hemisphere for each of the mutually overlapping 15-year periods changes its structure from a zonally-symmetric pattern similar to the Arctic Oscillation (AO) to a wavier pattern representing the AL-IL seesaw. The leading wintertime EOF for any of the 15-year periods based either on the observation or simulation appears to be represented as a linear combination of the AO and AL-IL seesaw.
It is argued that the AO (more precisely, the annular mode) and the AL-IL seesaw may be two of the fundamental patterns of the interannual tropospheric variability in the wintertime Northern Hemisphere, both of which can be generated by internal dynamics of the atmosphere. The remote influence of ENSO does not set a necessary condition for the seesaw formation. It is also argued that the midlatitude air-sea coupling might be important in reproducing the observed features. It will be discussed in the presentation whether or not the observed interdecadal modulation in the seesaw is attributable to any of the corresponding tendencies in the basic state.
Supplementary URL: