Large-scale modes of atmospheric variability. Part I. Statistical analysis and hydrodynamic modeling

摘要

Studied is the regional predictability of the low-frequency variability of the atmosphere and of the related patterns of atmospheric circulation for monthly and seasonal time periods in the Northern Hemisphere. Used is the SL-AV global semi-Lagrangian model worked out in the Institute of Numerical Mathematics of Russian Academy of Sciences jointly with the Hydrometcenter of Russia. It is demonstrated using the factor analysis that the model simulates successfully the first principal modes of variability explaining 85-90% of cumulative variance. Teleconnection indices which enable identifying the patterns of zonal flow and blocking, are used as the quantitative characteristics of low-frequency variability. In the framework of diagnostic verification, the useful signal was not revealed in the results of hydrodynamic modeling in the integration intervals of the second, third, and fourth months. The quality of forecasts at the monthly and seasonal time scales vary significantly depending on the region and season. Statistically significant correlations (in the limits of the 5-% significance level) between the actual and prognostic data are obtained for the regions of Northern Eurasia and North America only. No useful information was obtained in the forecasts of large-scale patterns over the oceans. The level of practical instability increases from winter to summer in all cases. The conclusion is made on the advisability of using the obtained results for monthly and seasonal weather forecasting.