Estimation of river high flow discharges using friction-slope method and hybrid models

Abstract

Abstract Severe flood is considered as one of the most important hydro-geomorphic hazards in catchment and urban scales. The correct estimation of flood flow in rivers is an important issue and plays a significant role in the optimal use of water resources, operation of dam reservoirs, and the design and planning of flood control projects. Despite the advances in hydrological models for flood prediction, these models suffer from many drawbacks and still have errors for the prediction. Hydraulic models also face with many uncertainties especially in riverbed Manning roughness coefficient and energy slope of the river. In this study, flood discharge estimation based on a new method of Friction-Slope or parameter α has been discussed for three alluvial rivers located in Golestan province in Iran. The calculation method is based on the Manning formula and takes into account the effects of energy slope of the river, as well as Manning roughness coefficient of the riverbed. For calculation of parameter α based on the simple and frequently available input variables of the river cross section (flow depth, flow area and hydraulic radius), hybrid methods of Particle Swarm Optimization-Support Vector Regression (SVR-PSO), Grey Wolf Optimization- Support Vector Regression (SVR- GWO), and Response Surface Method-Support Vector Regression (SVR-RSM) have been used and finally river flow discharges have been calculated and compared with the measured data. The optimum structures of hybrid models were determined using statistical evaluation criteria such as coefficient of determination (R²), mean absolute error (MAE), root mean square error (RMSE) and conformity index (d). The results showed that the SVR-RSM model has the best accuracy with R2=0.97, MAE=0.22, RMSE=1.66 and d=0.99 in the test phase. After determining parameter α using the RSM-SVR model, the river flow rates were calculated and compared with the observed values.