The Forecast of the Runoff of Georgian Mountain Rivers under Conditions of the Sparse Observational Network

Ts. Z. Basilashvili

Landscape and climate conditions of mountain rivers lead to high and long-lasting floods that sometimes result in ecological deformation and cause immeasurable damage to the environment and enormous economic loss for the country. Therefore, their forecast is of high scientific and economic importance. For the above reasons, a prognostic method based on the statistical model was worked out. The optimum system of predictors is selected from their potential set using various mathematical crite¬ria. By means of the direct and backward development of multi-factor dependences, the following regu¬larities are investigated simultaneously: the decrease in the number of predictors, the increase in the lead time and accuracy, and the possibility of ignoring accompanying factors that arise in the corresponding lead time. Different prognostic equations taking account of the differences in data, the lead time and ac¬curacy are derived for each hydrological section and region.