Using a Jump-Diffusion Process to Estimate Return-Period Flows

To estimate return-period flows in hydrology, a model based on a jump-diffusion process is used. In addition to the four model parameters, the dispersion parameter is also estimated for periods of rapid increase or decrease in flow. Simulations of the stochastic process are then performed. Using these simulations, the probabilities of exceeding significant flow values are estimated. Return-period flow estimates are obtained by interpolation or extrapolation, and are compared with the corresponding values calculated by hydrologists for the Delaware River in the United States.