The design storms developed from IDF curves are not representative real storm events. This is because the method of analyzing rainfall to obtain the IDF curves is independent of the real storm events. Each annual maximum volume for the 5, 10, 15, 30, 60, 120, 360, 720- and 1440-minute duration may come from a different storm event. This is demonstrated in the figure below. For the series of storm hyetographs shown, the maximum 5 min. volume occurred in event 4, and the maximum 10 min. volume in event 2. For time periods such as 12 hrs. and 24-hr., the maximum volume may come from more than one storm event.


In a real storm event, the maximum average intensities for fixed duration within the event have different frequencies. Real storm events do not have a single frequency. For example, event 4 has a return period of 2 yrs. for the 5 min. duration, a return period of 1 year for the 10-minute duration, and a 2-year return period for the average intensity of the total event duration. The frequency of average intensity for different duration in some storms may vary from 1 yr. to 50 years. Design storms on the other hand are developed so that the maximum average intensity for each fixed duration in the design storm has the same frequency. Since design storm are developed from IDF curves, they contain the maximum average intensities from different storm events. Design storms are therefore models of real storms and their validity requires testing.