Computation of probable maximum precipitation and its uncertainty

Abstract

Probable Maximum Precipitation PMP is used for estimating Probable Maximum Flood PMF which in turn is used for design of major hydraulic structures such as dams and spillways flood protection works and nuclear power plants One of the commonly used methods for estimating PMP is the statistical method also called Hershfield method that entails computation of frequency factor adjustment of the frequency factor construction of an enveloping curve of the frequency factor estimation of PMP choosing a probability distribution of PMP and determination of the return period of PMP There are however uncertainties associated with the PMP values estimated using the statistical method This study determined the PMP values for different durations using the statistical method with data from the Brazos River basin Texas It was found that significant uncertainty in the PMP estimates can occur with the use of enveloping curve of the frequency factor and the number of stations involved in its construction Hershfield rsquo s curve yielded higher frequency factor values by for hour duration by for hour duration and by for hour duration In comparison with basin specific values the PMP values from the Hershfield enveloping curve were higher for hour duration for hour duration and for hour duration For most of the Brazos River basin the return period of the PMP values was in the range of to years which was less than the range of to years reported in HMR showing the degree of risk associated with the PMP values Therefore a basin specific enveloping curve is suggested From commonly used statistical distributions and goodness of fit tests the Burr Type XII distribution was found to be the best frequency distribution for describing PMP It was observed that the return period obtained from the Burr type XII frequency distribution was not significantly higher than that obtained from the hydreometeorological reports HMRs of National Weather Service and other studies