Reliability of different depth‐duration‐frequency equations for estimating short‐duration design storms

[1] The design storm depth is defined as the rainfall depth that is expected to occur in the considered site for a given probability of occurrence and a given duration of the storm. It is generally estimated by using parametric depth-duration-frequency (DDF) equations. In particular, this article considers the frequent case when one needs to estimate the design storm depth for very short durations (typically from 5 to 45 min) while observed rainfall extremes for calibrating the DDF parameters are only available for longer durations. In this case the DDF curves are generally extrapolated below the range of durations that were used for estimating their parameters. It is well known that this procedure induces estimation errors that depend on the type of DDF equation that is used. The aim of this study is to test the capability of seven different DDF curves characterized by two or three parameters to provide an estimate of the design rainfall for storm durations shorter than 1 hour, when their parameterization is performed by using data referred to longer storms. The results point out that a proper choice of the DDF curve may improve the reliability of the design storm depth estimation significantly. In particular, DDF curves with three parameters provide a conservative estimation with an average relative error which never exceeded 20% for all the storm duration considered here.