Selection of design storms-time resolution considerations

The selection of actual historical storms for drainage system analysis is presented as an alternative to theoretical design storms. Selection considerations include rain gauge data quality, correcting rain gauge errors, as well as choosing long-enough and the right data subseries. A sample case study is presented, showing the statistical analysis of computed flows replacing the traditional analysis of rainfall data. An analysis of rainfall data time step sensitivity with several state-of-the-art runoff models demonstrates the different response characteristics. Time step size can be an important limiting factor for accepting or dismissing available data. Sensitivity appears to depend on model structure and complexity, as well as on the nature and variability of the storm hyetograph.