ABSTRACT
Risk analysis of extreme hydrological events should be emphasized critically in water resources management. Due to the obviously changing environment and urbanization, the stationarity assumption of extreme events for performing the hydrological frequency analysis to make a systematic risk assessment is challenging and problematic. We investigate nonstationarity and trends in the annual precipitation extremes in six selected gauges in Haihe River Basin of China. This paper is aimed at make a deep comparison of three models in terms of trend analysis, goodness of fit, extreme rainfall quantiles corresponding to different return periods. This results denote the importance of considering nonstationarity when assessing the return period and hydrological risk. It is concluded that nonstationary risk analysis of annual maximum rainfall series can be necessary to water resources management coping with climate change.
