STATISTICAL HYDRODYNAMIC LONG-RANGE FORECASTING OF METEOROLOGICAL FIELDS WITH THE MGO MODEL. 1. STAGES OF IMPROVEMENT OF THE METHOD AND FORECAST PROCEDURES

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

A statistical hydrodynamic method of forecasting meteorological fields for up to a month is considered. The method is based on the application of a global atmospheric general circulation model (T30L14, MGO) and the ideology of the ensemble forecast. Initial fields are set from results of the objective analysis of the Hydrometeorological Center of Russia, while the ensemble of 11 initial states is formed from initial fields by means of specially simulated disturbances. Boundary conditions close to actual ones (the observed initial sea surface temperature and calculated soil humidity) are taken into account. Results of integration for 28 days are averaged over the ensemble and over the forecast periods: 1, 1—2, and 1—4 weeks. Systematic errors are corrected. The main prognostic variables are the fields of the 500 hPa geopotential, air temperature at 850 hPa, and sea-level pressure (at 2.5 × 2.5º grid points for 40—90ºN), and also air temperature at 2 m from the CIS network stations. The forecast is made once a week (with a week overlapping). The method is realized on the PC Pentium-III and since 1 January 1999 has been functioning in the operational regime.