Abstract: Previous flood simulation methods only consider correlations among flood peak, volume and duration, but neglect the influence of flood characteristics on the shape of flood hydrograph. This paper describes a stochastic flood simulation method considering the shape of flood peak and its frequency by introducing peak shape coefficient and peak time. The copula function is used to construct a joint distribution function of these variables and simulates them though stochastic simulation. By calculating closeness degree, we can obtain the representative types of dimensionless flood hydrographs that are the closest to the stochastic simulations, and then achieve a complete flood hydrograph through combining the two functions and apply it to numerical simulations for flood control and dispatching. Case studies of the Dalong reservoir show that our method comprehensively considers the major factors that affect flood process. And the frequency of each simulated peak type is close to that of the measured flood, and the calculated risk rate of flood control is not much different from its real value. Thus the method is useful for formulation of reservoir dispatching schemes and risk analysis.

Key words: peak shape coefficient, peak time, flood hydrograph, stochastic simulation, risk analysis