Probabilistic Models of Extreme Flood Water Discharges in Rivers of Cisbaikalia

Abstract

This article considers the possibility of using probabilistic models to analyze the maximal flow discharges of rivers in order to obtain reliable calculated statistical characteristics for basins with poorly studied hydrological features. The research was performed by the example of Cisbaikalia, which is characterized by a flood regime of river flow. It is found that floods in the study area most often occur in summer (July−August), are associated with the climatic characteristics of the region, and are often destructive. The analysis of the maximal flow of rivers is based on data from the Roshydromet observation network. The series of maximal water discharges are checked for homogeneity and, in general, no disturbances in the steady state of runoff caused by climate changes are detected. A generalized distribution of extremes is proposed as the main probabilistic model; it is recommended to determine its parameters on the basis of the group analysis. The integrated approach has been applied for the first time; it combines conventional methods of hydrological calculations, which are most often used to refine the characteristics obtained for the runoff in the zone of extreme values: the apparatus for truncation of distributions; joint analysis of data; a reduction formula with the reduction of the drain modulus value not only to the area of 200 km², but also to the mean height of basins in the region; and the frequency probability method for estimation of obtained results. These methods are recommended by regulatory documents for discharge calculations and are most often individually used. The comprehensive approach described by the authors enables us to take into account the features of the runoff formation in the zone of extreme values and obtain more accurate values of characteristic quantiles of a given probability of excess for use in design on poorly studied rivers of the region.