STATISTICAL HYDRODYNAMIC LONG-RANGE FORECAST OF METEOROLOGICAL FIELDS WITH THE MGO MODEL. 2. OPERATIONAL TEST RESULTS AND PROSPECTS OF IMPROVING THE PROGNOSTIC SCHEME

V. P. Meleshko, V. M. Gavrilina, V. M. Mirvis, V. A. Matyugin, Yu. A. Pichugin, and S. V. Vavulin

Test results of a hydrodynamic and statistical method of the long-term forecast of meteorological variables developed at the Voeikov Main Geophysical Observatory (MGO) and based on the integration of equations of the global spectral atmospheric general circulation model T30L14 are described. The forecasts were made once a week for up to one month in advance from January 1999 through July 2001. They contained values of 500-hPa geopotential, 850-hPa air temperature, and sea-level air pressure at 2.5 × 2.5º grid points and also surface air temperature from the CIS network stations. The values were averaged over one week, one—two weeks, and one—four weeks. A comprehensive analysis of the quality of forecasts is made. For periods of one week and one—two weeks the method is shown to have an obvious advantage over persistence and climatic forecasts. The forecasts for up to one—four weeks are characterized by a stable positive correlation of prognostic and actual anomalies. With respect to the root-mean-square error, they are at the level of climatic forecasts. Prospects and the ways to improve the method are discussed.