METHODS OF LONG- AND MEDIUM-RANGE FORECASTING OF DATES OF THE MAXIMUM SPRING FLOOD LEVEL ON EUROPEAN RUSSIA'S RIVERS

B. M. Ginzburg, S. V. Borshch, N. D. Efremova, M. I. Silnitskaya, and K. N. Polyakova

Long-term characteristics of the dates of the beginning of flood rise and of the occurrence of its maximum level under current climate conditions on rivers in European Russia are determined. It is found that a close relation (r = 0.80—0.90) exists between the flood onset and breakup dates, while the correlation between breakup and maximum level dates is only significant (r = 0.50—0.80). The dates of flood peaks are under strong influence of a space distribution of the sea surface temperature (SST) in the North Atlantic and northwestern Pacific (R = 0.68—0.73) in winter months. It is possible to find complex relationships between the dates of flood peaks and characteristics (EOF expansion coefficients) of SST fields and H100 and H500 geopotential fields over significant zones of the Northern Hemisphere in January—March (S/σ = 0.73 ± 3). A new forecasting method has been developed, which is a background long-term forecast and a medium-term forecast (correcting the long-term forecast) of the dates of maximum spring flood on major free-flow rivers in European Russia. The success of this long-term forecast based on the above relationships is about 75%, and for early events, it attains 80%. The corrections of the long-term forecast on the basis of the joint consideration of the ten-day forecast of air temperature and SST in preceding months have an average lead time of 26 days, which raises the total accuracy of the forecasts to 88 ± 3%.