Variability in Design Flood Flows from Alternative Rainfall Temporal Patterns

Design flood estimation remains a problem for many engineering hydrologists. Advice is required regarding design flood characteristics for design of culverts and bridges, urban drainage systems, flood mitigation levees, dam spillways, and many other situations. The important flood characteristic depends on the nature of the problem, but typically is either the peak flow rate, peak level, hydrograph rate of rise, flood volume, or system failure. As discussed by Ball et al. [2], a common approach for estimation of the peak flow is application of catchment modelling systems with the frequency of rainfall intensity transferred to the predicted peak flow rate. This approach assumes the temporal variability of rainfall does not influence the frequency transferral. Using rainfall temporal patterns generated using the approach of Varga et al. [12], 60 alternative non-dimensional patterns were generated. These alternative temporal patterns were combined with the Intensity-Frequency-Duration information at Sydney to generate storm bursts over the full range of frequencies, namely AEPs of 50% to 1%. The resultant storm bursts were applied to a SWMM based model of the Centennial Park catchment in Sydney Australia. Predicted flows for the various frequencies were analysed to ascertain the variability in the predicted peak flow arising from the temporal pattern of the rainfall and hence the validity of the assumption that the temporal pattern of rainfall does not influence the transferral of the rainfall frequency to the predicted flow frequency. Presented in this paper are the results of an investigation into the validity of this assumption