Run Hydrographs for Prediction of Flood Hydrographs

A bivariate lognormal probability distribution in flood peaks and flood volumes describes families of flood hydrographs, each family with a desired return period. Truncated series of floods are used to estimate the underlying population parameters. The method provides more information on floods than conventional methods, and the theory is illustrated with a worked example. Run hydrograph theory was developed and demonstrated for application on fully-gaged catchments. The method allows for the synthesis of the whole family of design hydrographs for each desired return period. In the development of the theory, hydrographs were analyzed as they occur in the record; thus the sample of hydrographs contains many composite hydrographs. This should result in realistic design hydrographs for practical applications. In the estimation of the necessary parameters it was demonstrated how straightforward it can be to calculate the underlying probability distribution parameters from a truncated sample. The run hydrograph technique may also be amenable towards extrapolation to ungaged catchments.