Abstract:

This paper studies the application of partial L-moments to frequency analysis for estimation of design flood with large return period to reduce the disturbances of small floods in the flood series. Monte-Carlo experiment was conducted to examine the statistical properties of partial L-moments. This paper analyzes the imitative effect and evaluates the errors in estimation of design flood by using the annual maximum flow series at five hydrological stations in Shaanxi province. The results show that the partial L-moments method has good statistical properties. In comparison with traditional L-moments, this method can better describe the data samples in flood analysis and improve the estimation accuracy of design flood. Thus, it is a reasonable and effective method for flood frequency analysis.