Power spectral analysis of the principal component timeseries reveals that both the PNA and NAO can be regarded as first order Markov processes with a decorrelation timescale of 7.7 and 9.5 days, respectively. Furthermore, the analysis also indicates that the PNA and NAO lifecycles are complete within in approximately two weeks, rather than several months as suggested by those studies that use longer time-averaged data.
The diagnostic and modeling calculations show that the growth of the two upstream PNA anomaly centers arises from barotropic energy conversion from the zonally asymmetric climatological flow, and the two downstream PNA anomaly centers grow through linear dispersion. Divergence is found to play a crucial role by maintaining the PNA anomaly in a quasi-fixed position, which further enhances anomaly growth. The characteristics of the PNA decay are found to be consistent with Ekman pumping. The transient eddy vorticity fluxes are found to play a less important role than previous studies have suggested.
The growth of the NAO anomaly is found to be driven by eddy fluxes. Both the high-frequency (period <10 days) and low-frequency (period >10 days) transient eddy fluxes contribute to the NAO growth. Once the NAO anomaly reaches its largest amplitude, the eddy forcing quickly becomes negligible, and the NAO decays primarily through linear dispersion.
These results illustrate many important differences between the PNA and NAO lifecycles. Perhaps, most strikingly, they suggest that the PNA lifecycle may be viewed as a linear initial value problem, whereas the NAO lifecycle can be interpreted as the response to nonlinear, transient eddy forcing.
10>