ESTIMATION OF SUSPENDED SEDIMENT LOAD BY MULTI LINEAR REGRESSION ANALYSIS

Abstract: Suspended sediment load (SSL) is defined as the rate of sediment transported by a running water stream. It is essential to have an idea about the rate of sediment transported for the solution of river engineering problems. There is no standard specified method for estimating SSL. In last decades, researchers often have been using machine learning methods in order to predict SSL with increasing the development of the computer technology. In this study, the monthly flow rate and suspended sediment load (Qs) of Karamenderes Stream in Turkey between the years of 1996-2004 were estimated by using multi linear regression (MLR) analysis on SPSS program is studied. The load of the SSL is determined by using depending parameters of water temperature (Co) and flow discharge (m3/s) and corresponding independing values of suspended sediment load. Performance of the MLR model is measured by using coefficient of determination (R2). In multi linear regression analysis, most effecting parameters of the prediction of the SSL are flow discharge and temperature according to order of importance. Data of model are predicted with 85 % approximation.