COMPARATIVE ANALYSIS OF METHODS FOR FORECASTING MAXIMUM LEVELS AND VOLUMES OF FLOOD RUNOFF OF A MOUNTAIN RIVER

Abstract

The study topicality relates to the analysis of available methods for assessing snow reserves in river basins and forecasting floods. To predict runoff volume and maximum flood levels, statistical models resting on the correlation dependence of runoff parameters on snow reserves or multiple linear regression equations are usually used in the absence of sufficient hydrometeorological information. An important point is that there are no justified comparisons of traditional forecasting methods based on calculations of winter precipitation amount (snow reserves) and statistical models of multiple regression. The aim of the study is to carry out the comparative analysis of methods for calculating winter precipitation amounts in the basin by means of traditionally applied altitude dependencies and the author's estimation method using orographic correction to the velocity of vertical movements of air masses; to carry out the comparative analysis of traditional statistical forecasting methods and models of multiple linear regression by the example of the Amyl River basin. Methods: comprehensive geographical and hydrometeorological analysis; dependence establishment based on long-term hydrometeorological observations using methods of mathematical statistics (correlation and regression analysis); snow reserves simulation with the use of orographic correction to the velocity of vertical movements. Results. By altitude dependencies and the author’s assessment considering orographic correction to the velocity of vertical movements of air masses, two different approaches to estimate of winter precipitation amounts suggest similar statistical dependences with close values of determination coefficients. Models for predicting runoff volumes and maximum flood stages based on pair correlation and multiple regression analysis were developed. The comparative analysis of traditional methods for forecasting runoff layers based on calculations of winter precipitation amount (snow reserves) and statistical models of multiple regression also showed similar results. The applied one-factor dependencies turned out to be the best in forecasting snowmelt-induced maximum levels.

Keywords

• zapadny sayan
• tuba river
• amyl river basin
• winter precipitation
• snow reserves
• altitude dependence
• orographic correction to the velocity of vertical movements
• statistical model of multiple regression