Estimating Suspended Load by Linear Piecewise Regressions

Sediments are materials (Loads and Bed-Loads) that are transported by river flows.Sediment rating curve and power models are usually used for estimating suspended-loads data. Power modelsin large floods (extrapolation) are not suitable because of their exponential growth. Piecewise models are presented here to overcome this behavior. Piecewise regression models are applied for 18 basins suspended- Loads data. These models have one (or two) nods. Adjusted r-squares of these models are more than 0.5 in all cases. F statistics are between 46 and 1975 for all 18 models. P-values of all parameters are much less than 5% that shows necessity of parameters. Tables of ANOVA and diagnostics of residuals have been done; they showed goodness of fits. Results show that Piecewise regression models have good behavior.