NONPARAMETRIC-ESTIMATION OF LOW-FLOW FREQUENCIES

A new nonparametric method is proposed for low-flow frequency analysis . Low-flow series often contain years with zero values. The occurrence of zero and nonzero events are treated by first estimating the probab ility of zero flows using a plotting formula, and then fitting the non zero events by nonparametric method. A Monte Carlo-simulation experime nt is performed to compare the proposed nonparametric method with two parametric distributions, namely, log-Pearson Type III and Weibull. Re sults show that the nonparametric method gives more accurate results t han currently used parametric methods. It is also concluded that the n onparametric method is uniform because it does not require a priori sp ecification of any parametric distribution.