Determination of maximum and mean discharges of Black Sea Region streams with basin physiographic parameters

Average flow and maximum flow in streams are two important flow characteristics used in the design of water structures. These flow rates can be obtained from stream gauging station data. However, not all streams have flow monitoring or do not have sufficient data. Flow characteristics of basins that do not have discharge data are generally transferred using the ratios between the flow rates in neighboring basins and basin areas. This study aims to determine the most appropriate linear regression models that can be used to calculate the long-term average flow (Qave) and instantaneous maximum flow (Qmax) values for the Black Sea region rivers. For this purpose, the basin physiographic parameters and land use rates obtained from the Digital Elevation Model (DEM) of the basins of the 35 stream gauging stations, which have been observed for at least 20 years, were used as the data of the regression analyzes with their original values and inverses of the cube roots. Basin physiographic parameters were obtained from the Digital Elevation Model of 1 angular second. Strahler bifurcation value, basin area, basin circumference, circularity ratio, shape factor, average slope, sinusoidity, compactness coefficient, basin depth, relief ratio, elongation ratio and drainage density were taken into consideration as basin physiographic parameters. To represent land use, the catchment forest ratio, which is thought to have the most impact on the flow, was used. Relationships between flow characteristics and parameters were investigated by multiple linear regression analysis. In the regression analysis, first the original values of the parameters were used, then the same analysis was repeated with the inverses of the cube roots for linearization. The best regression models were determined with the functions in the libraries of the R program. The performances of the models were evaluated with coefficient of determination, adjusted coefficient of determination, Akaike Information Criterion and Mallows coefficient. It has been seen that taking the inverses of the cube roots of the data increases the success of the regression models.