STABILITY OF ATMOSPHERIC PROCESSES AND AN A PRIORI SKILL ASSESSMENT OF HYDRODYNAMIC FORECASTS

A. V. Murav'ev and I. A. Kulikova

Stability and predictability of atmospheric processes are discussed according to a priori skill assessments of hydrodynamic forecasts with lead time over midranges. An overview of foreign publications is given. Results of numerical experiments with a global T41L15 general atmospheric circulation model are presented. Hydrodynamic forecasts were performed for 30 days ahead every 10 days with initial data from December 1996 to August 1998. Quality forecasts were made on the basis of ARMA models and nonparametric discriminant analysis. The predictors were the root-mean-square difference between three consequent 500-hPa height fields, the anomaly or amplitude of the 500-hPa height field, and baroclinic instability indices. It is shown that these parameters contain much information on hydrodynamic forecast quality for the 10-day period.