JWSS - Journal of Water and Soil Science

Information on suspended sediment variation in times of flood is important in management of water resources, particularly management of basins, and in investigation of the causes of erosion. The relationship between discharge and suspended sediment concentration during floods is not similar and homogeneous for different reasons such as precipitation variety, discharge rate and sources of sediment and production of hysteretic loops. In this study, the instantaneous values of suspended load were simulated using genetic programming and regression methods. By comparing the two models, Genetic programming model was selected as the better one with the mean square error and determination coefficient of 0.8 and 0.5, respectively. Then based on this model, loops of suspended hysteretic load were drawn for the six events recorded in the period of 1387-1383. This resulted in 4 linear and 2 clockwise hysteretic loops for the river suspended sediment. Identifying various hysteretic loops is effective in determination of relative contributions of processes to production and transfer of sediment including amount and intensity of precipitation, flow rate and previous moisture conditions of watershed. The results showed that the clockwise hysteretic loops occurred usually in high precipitation and discharge, and linear hysteretic loops in spring because of low intensity precipitation.

Non-linear Muskingum model is an efficient method for flood routing. However, the efficiency of this method is influenced by three applied parameters. Therefore, efficiency assessment of Imperialist Competition Algorithm (ICA) to evaluate optimum parameters of non-linear Muskingum model was addressed in this study. In addition to ICA, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) were also used to find an available criterion to verify ICA. In this regard, ICA was applied for Wilson flood routing then, routing of two flood events of DoAab Samsami River was investigated. In case of Wilson flood, the target function was considered as the sum of squared deviation (SSQ) of observed and calculated dischargem. Routing two other floods, in addition to SSQ, another target function was also considered as the sum of absolute deviations of observed and calculated discharge. For the first floodwater based on SSQ, GA indicated the best performance however, ICA was in the first place, based on SAD. For the second floodwater, based on both target functions, ICA indicated a better operation. According to the obtained results, it can be said that ICA could be recommended as an appropriate method to evaluate the parameters of Muskingum non-linear model.