Hybrid Stochastic Model for Daily Flows Simulation in Semiarid Climates

An innovative stochastic model for the purpose of daily flows simulations is presented in this paper. This model is particularly well suited for the generation of daily river flows in semiarid climates. The previous models have not preserved basic features of daily flow hydrographs, such as peaks and recessions, and have failed to reproduce the long low water and short flood periods due to the temporal distribution of precipitation in semiarid climates such as Sahelian regions. This hybrid stochastic model uses a product model that simulates an intermittent impulse series. The impulse magnitude is the flow increment from one day to the next. The intermittent impulse series enables the definition over time of alternate rising and falling limbs. On the rising limbs, the generated impulse magnitudes help to produce the daily flows. However, on the falling limbs, the flows are produced by a deterministic model using a time-varying recession coefficient that is, for the first time, introduced in stochastic modeling of daily flows. Application of this model for daily flows simulation at the Bakel Station on the Senegal River (West Africa) has provided good results as shown by the analysis of the synthetic daily flow series.