Abstract:

Rainfall and runoff are two important hydrological elements in a basin, which have stochastic characteristics. With the intensification of human activities and global climate change, the relationship between rainfall and runoff in a basin is becoming more and more complex. Based on the advantages of Copula function in describing the correlation of stochastic variables, this paper firstly introduces the nonparametric kernel density estimation method to estimate the marginal distribution of rainfall and runoff variables in the Fuchunjiang river basin. Then, the binary Copula function is used to construct the joint distribution model. The simulation effects of the marginal distribution and the joint distribution are tested by the root mean square error and the Euclidean distance. The results show that the Gaussian kernel function has a good simulation effect on the marginal distribution, and the Gumbel-Copula function has a high degree of fitting on the joint distribution. There is an upper tail correlation between annual rainfall and annual runoff in the Fuchunjiang river basin, and the probability of simultaneous maximum is 75.83%. The results can provide reference for the risk response of water disasters, water resources scheduling management and water engineering planning and design, and have important theoretical and practical significance.

Key words: Copula function, kernel density estimation, correlation analysis, rainfall, runoff