Detention rockfill dams are an easy and common tool for f lood control. Due to their coarse pores, the flow in void spaces is turbulent and non-Darcy. Diff erent relationships introduced by researchers are used to define the hydraulics of the f low within the rockf ill materials. The present research is aimed at gaining a better understanding of the diff er-ence among these relationships and the sources of uncertainty associated with the diff er-ent parameters of each of the relationships. To

more »

... amine the importance of various f actors on the uncertainty of the outf low hydrograph, sensitivity analysis was conducted. For this purpose, a rockf ill mass was provided, f if teen random samples of the mass selected, and then the physical characteristics of the material were measured or estimated. Also, some flood routing tests have been conducted. In these tests a physical model of a dam was installed and downstream water level was measured f or diff erent outflow rates. While the downstream water level was considered as certain variable but other parameters were seen as stochastic (stochastic parameters are considered as random variables) and outf low discharge as an output uncertain parameter. Uncertainty analysis has been conducted f or different points of the outflow hydrograph by employing available methods. The results show that the Samani et al. and McCorocoudale et al. relationships have the lowest and highest uncertainty, respectively. The sensitivity analysis demonstrates diff erent levels of sensitivity accompanied each of the relationship parameters which results in diff erent ef-f ects on the total uncertainty of the relationships. IN TRO DUC TIO N Uncertainties may arise due to natural variations in the phenomenon being considered , or to an incompleteness of our understanding. Uncertainties may also arise from the inaccurate characterization of important parameters or variables. Hence, engineering practice is frequently associated with decision making under uncertainty. T he physical or numerical models, developed and used to simulate natural phenomena, are often in reality probabilistic, and hence, subject to analysis by rules of probability theory. Identifying the components of uncertainty related the physical phenomenon and quantifying them, can therefore improve decision making and the results (Haung, 1986; Mercer, 1975). One of the common and most economic methods for flood mitigation used in watershed management is rockfill dams. The fact that flood mitigation through rockfill dams is an uncertain phenomenon, raises questions about the reliability and credibility of the relationships involved. As this type of dam consists of coarse particles, the flow deviates from Darcy's law resulting in turbulence in the void spaces. This means that the relationship between the flow velocity, V, and its hydraulic gradient, i, is a nonlin-ear one. Different researchers have proposed different nonlinear relationships that give various outputs.