THE GROWTH AND DECAY OF LOW-FREQUENCY ANOMALIES IN A GCM

The temporal evolution of a regional-scale persistent low-frequency an omaly is examined with data from a 2100-day perpetual January general circulation model. The persistent episodes are determined with an obje ctive analysis of the low-pass (>10 day) 350-mb streamfunction field t hat uses both pattern correlations and empirical orthogonal function ( EOF) analysis. The composite evolution of each term in the streamfunct ion tendency equation is calculated relative to the onset day (the fir st day of the persistent episode). By projecting each term in the stre amfunction tendency equation onto an EOF (the EOF is associated with a particular low-frequency anomaly), the contribution of these terms to ward the tendency of the corresponding principal component can be obta ined. It is found that the sum of the linear terms dominates during mo st of the growth and the decay of the low-frequency anomaly. The linea r term that accounts for the growth and maintenance of the low-frequen cy anomaly is the interaction between the anomaly and the time-mean zo nally asymmetric flow. After the anomaly attains sufficient amplitude, its decay is accomplished through the divergence term. For one phase of the EOF, it is found that the high-frequency transients prolong the anomaly, whereas in the other phase they do not. Implications of this study for examining monthly averaged anomalies are also discussed.