Abstract: To study the impacts of climate change on the occurrence frequency of flood characteristics in Bejiang River basin, the outputs of two scenarios (RCP4.5 and RCP8.5) based on BCC_CSM1.1 are statistically downscaled by Quantile mapping (QM) and used as input of SWAT model to simulate streamflow during historical (1965-2010) and future (2030-2064, 2065-2099) periods. Univariate analysis and bivariate joint analysis based on Copula are used to analyze annual maximum flood peak flow (Q) and annual maximum seven-day flood volume (W). The results show that climate change has heavier impacts on Q and W with larger return period exceeding 50 years, except from the W of 2065—2099 under RCP8.5. Climate change under RCP4.5 has heavier impacts on Q and W than that under RCP8.5. The impact of climate change on the Q is greater than that on the W under two scenarios. The design floods of the bivariate joint analysis is safer than that of the univariate analysis. The study incorporates climate change and bivariate joint analysis, and of referential value for flood risk assessment and management under varying circumstances.

Key words: flood peak volume, climate change, bivariate analysis, Copula function, frequency analysis, Beijiang River basin