MONTHLY METEOROLOGICAL FORECASTING WITH THE MGO HYDRODYNAMIC-STATISTICAL METHOD. I. DETERMINISTIC FORECAST: RESULTS FROM 26-YEAR SERIES OF TESTING

V. M. Mirvis, V. P. Meleshko, V. M. Gavrilina, V. A. Matyugin, and T. Yu. L'vova

A new version of the hydrodynamic-statistical method of meteorological forecasting with a time lead of up to one month is described. The method is based on the MGO global spectral atmospheric general circulation model (T42L14) and an ensemble approach. Results are analyzed of multiyear (from 1979 through present) series of hindcasts using the NCEP/NCAR reanalysis data and of a 4.5 year series of forecasts using operational objective analysis data from the Hydrometeorological Research Center of Russia. The emphasis is on forecasts of surface air temperature for 70 stations located in Russia, CIS, and Baltic. For the first time, skill scores are given for predicted monthly sums of precipitation. Dependence is considered of the quality of air temperature forecasts on the season and the territory and on the lead time. Expediency is demonstrated of weekly structuring of the forecast within a month with at least once in two weeks updating.