Abstract: Under the combined action of multiple variables of flood, hydraulic projects are subjected to risks of inconsistent flood control capaciy caused by flood control parameters calculated from existing multivariate return periods. Hence, the definition and calculation of multivariate return period has become a difficulty in flood frequency analysis. In view of this, a new definition for multivariate return period is proposed in this paper, that is, flood control return period. Flood control return period is defined as the average time between two successive events (event is defined as the flood control design parameter being exceeded). The coupling relationship between flood combinations and flood regulation are taken into consideration. The proposed return period is compared with traditional bivariate return periods in terms of definition and risk area delimitation. With Guiping shipping hub as a case study, the differences between the proposed flood control return period and the traditional bivariate return periods in flood control design are calculated and examined by constructing the joint distribution of floods in the mainstream and tributary using the Copula function. The rationality and reliability of flood control return period are verified. The results manifest that flood control capacity can be consistent with flood control standard by using the proposed return period. Overestimating flood risks and increasing of engineering cost, or underestimating flood risks and increasing of flood damage can both be avoided.

Key words: multivariate flood, bivariate return period, flood control return period, Copula function, risk area, design combination value