The genesis and predictability of persistent Pacific-North American anomalies in a model atmosphere

The setup process of Pacific-North American (PNA) pattern anomalies that la st more than 10 days and the role played therein by synoptic-scale transien ts are investigated using a T21, 3-level quasi-geostrophic model. As there is no time-dependent forcing in the model, the low-frequency PNA anomalies are generated entirely by the internal dynamics. From a 300-winter integrat ion, 100 cases of positive PNA anomalies and 118 cases of negative PNA anom alies lasting at least 10 days are identified. The PNA composites reveal th at 5 days before the setup of a positive (negative) PNA anomaly there is a negative (positive) height anomaly over Canada. This anomaly moves westward and stops over the North Pacific. It then intensifies and a wave train dev elops downstream. The role played by the synoptic-scale transients is explo red by prediction experiments starling from day - 5 (5 days before the appe arance of a PNA anomaly). A 5-day average centered at day -5 is used to rem ove the synoptic-scale transients from the initial state. The results indic ate that even without these transients in the initial conditions, the setup process of a positive PNA anomaly can be reproduced quite well. On the oth er hand, the setup of a negative PNA pattern cannot be predicted if the syn optic-scale transients are absent from the initial conditions. In this case , the pattern evolves systematically, but differently from the way it evolv es in the control run.