Potential of Mean Daily Flows for Improving Peak Flow Quantiles Estimation

Design and operation of hydraulic infrastructures require the assessment of quantiles of peak flow discharge, which is generally carried out through the fitting of a parametric probability function to an observed series of peak flows. However, in many cases, peak flow series are only partially available, or not available at all, while records of mean daily flows are generally more common. Indeed the mean daily flow can be considered as a right censored data with respect to peak flows, being the former always less or equal than the latter. Therefore, techniques for parameter estimation for censored data, usually applied for paleo-flood analysis, are potentially applicable. In the paper, we investigate whether including information from available mean daily flows series may increase the reliability of the estimated peak flow quantiles. In particular, the analysis is carried out with reference to three gauged stations within the Simeto River Basin, the largest watershed in Sicily, Italy, for which contemporaneous observations of peak flows and mean daily flows are available. Several combinations of series of peak flows and mean daily flows of different lengths are analysed by means of a resampling approach.