River runoff modeling in the European territory of Russia

Abstract

The paper describes river runoff modeling for a plains region of the European territory of Russia (ETR), as well as a prediction for ungauged drainage basins. The study of river runoff is one of the key research objectives in determining the patterns of sediment yield formation. Among many other zonal factors, river runoff is considered to be the main factor in sediment yield formation in a humid climate. In this study, modeling results for the entire European territory of Russia and various landscape zones are presented via the use of multiple regression methods. Multiple regression methods do not require the mathematical description of the main physical processes of runoff formation in terms of their spatial heterogeneity. At the same time, such methods can be distinguished by their simplicity in terms of determining parameters and providing clear interpretations of the results. The research methodology in this work is based on a drainage basin approach. Initial data for the river runoff and its formation factors are presented in the open-access geoinformation database “Drainage basins of the European territory of Russia”, which has been created earlier by the authors. The river runoff geodatabase was formed with results from 1440 gauging stations. The independent variables, such as the relief morphometric characteristics, climatic indicators reflecting average values, scale, seasonal variations, extreme values of temperature and precipitation, percentage of forest and swamp cover, plowing, percentage of meadows, assessment of the anthropogenic impact on the drainage basin, geographical coordinates of the centroid, prevailing soil type, type of soil-forming rock, and class of pre-Quaternary deposits are used for modeling here. Data processing and model development is conducted using the R software environment. Models obtained by linear and nonlinear methods explain about 85–88% of data variability and are well interpreted in terms of the water balance equation. It is found here that the most significant predictors in the model are annual precipitation, the sum of the active temperatures (characterizing runoff losses via evaporation), average slope gradient, and the forest cover of the catchment. For Environmental Resources Management, it is required that data for river runoff are collected at the local (municipal) level. The results for the extrapolation of the river runoff values to ungauged river basins in a plains region of the European territory of Russia are presented here. Calculations of predicted values for the river runoff are given based on the obtained discharge per unit area logarithm model. The model and its cartographic representation reflect the patterns of the spatial distribution of river runoff for the level of spatial detail accepted in the study. The methods applied in this study and the results obtained could be used for similar studies of plains territories across the world.