Frequency Analysis by Bootstrap, SIR Algorithm, and Parametric Method

The frequency analysis is one of the most useful statistical methods in water resources engineering fields. However, there are some problems such as the need of many samples, the change of estimated values according to the parameter estimation method, and the change of distribution type according to the number of measured events. So, in this study we used the Bootstrap method and SIR(Sampling Important Resampling) algorithm which are non-parametric methodologies as the alternatives of the parametric method for the estimation of rainfall and flood frequencies. Before applying Bootstrap method and SIR algorithm, data collection is an essential step. We estimated the rainfall frequency using the parametric, Bootstrap method, and SIR algorithm. Therefore the alternative methodologies may be very useful for the determination of the design values of hydrologic observations by the comparison with the design values by the parametric method. We also expect that the alternatives could compensate for the problems of parametric method.