Temporal patterns for design rainfall bursts at present are based on the Average Variability Method (AVM); an approach first proposed by Pilgrim and Cordery (1975) for estimating the temporal pattern of rainfall in the most intense burst of rainfall within a storm event.

This approach is known to result in storm bursts that have higher temporal correlations than exist in real storm events. This impacts on the estimation of flood flows due to the influence of the temporal pattern on the shape of they hydrograph; Ball (1994) showed that the rainfall temporal distribution significantly influenced the magnitude of the peak flow and more particularly the shape of they hydrograph. Furthermore, the validity of the assumption that an AVM pattern transforms the rainfall frequency to the flood frequency is unknown and untested.

Finally, there are many problems where the storm volume is significant and hence there is a need to consider not only the temporal pattern of the peak storm burst during an event but also the temporal pattern of rainfall prior to and post the peak burst of rainfall; a problem which has been discussed by Phillips (1984) and Rigby et al. (2003).