Abstract
Experiments on dynamical extended-range forecasting (DERF) conducted at the National Meteorological Center (NMC) are analysed for the relationships between skill of medium-extended-range forecasts over the Pacific-North America region and the fluctuations of the Pacific-North American (PNA) mode of low-frequency variability. To isolate the effects of the prominent El Niño-Southern Oscillation anomalies that prevailed during the DERF period, the performance of forecasts is evaluated separately for the North Pacific and the North Atlantic sectors. Distinct features are observed. Much better skill in both dynamical and persistence forecasts is found for the Pacific sector than the Atlantic sector.
The relationships between the polarity and amplitude of the PNA mode and the predictability of the prediction model are also investigated. The PNA circulation regime in the initial conditions as a predictor of forecast skill is contrasted in detail with the PNA mode in the forecasts. The statistical significance of the relationships is examined. The results indicate that the PNA mode in the forecasts is a better predictor for forecasting the forecast skill than the PNA mode in the initial conditions.