The generalized extreme-value (GEV) distribution has facilitated many applications in hydrology, used to model a wide variety of natural extreme events. Previous studies show that small-sample maximum-likelihood estimators parameters are unstable and demonstrates that absurd values of the GEV shape parameter can be generated. It is recommended that LH moments estimators, based on linear combinations of higher-order statistics, should be introduced for characterizing the upper part of distributions and larger events in data. However, there have been no computer packages for PC that model extreme events by LH moments. The objective of this paper was to present the modeling of hydrological extreme events by GEV distribution, using LH moments to estimate its parameters and the goodness-of-fit test proposed by Wang (1998) to evaluate the goodness-of-fit, both developed in Matlab. The results are presented for the estimation of the parameters of the GEV, the values coefficient of variation, skewness, kurtosis and goodness-of-fit test values, and fitting extreme flow observed in Paraná Rivers by GEV distribution.
frequency analysis; estimation; hypothesis testing; extreme events
