Using Volume Delivery Time to Identify Independent Partial Series Events

In practice, the annual series of streamflow peaks is generally preferred than the partial series for flood frequency analysis. Flood selection criteria for the partial series tend to be arbitrary and are limited in making allowances for catchment scale. This aspect appears to be a constraint to greater acceptance of the partial series approach. The aim of this paper is to define a scalable selection criterion that reduces ambiguity in flood selection by defining floods that exceed the daily average. An approach based on the volume delivery time (VDT), analogous to the tip interval time in tipping bucket raingauges, is described and tested for three rural catchments of various sizes in South East Queensland, Australia. The VDT approach produced discharge quantile estimates similar to the partial series based on the commonly-used monthly maxima except for minor, high frequency discharges at the small, more perennial catchment. A simplified approach based on average daily volume gave similar results to the VDT method.