Research on selection of prior distribution of Bayesian theory in hydrological uncertainty analysis

Bayesian theory plays an important role in hydrology uncertainty analysis. In the theory, how to choice the prior distribution correctly is a critical issue. At present, there is no good idea in prior distribution estimation; the only way is to accordance with the principle of non-uniform distribution of the information instead. In this paper, the sensitive parameters of Xin´anjiang model are used to analysis the influence of the prior to posterior distribution, under the reasonable range of parameters in taking several common distributions. The study can provide new reasonably ideas in prior distribution estimation for the future and can improve the Bayesian approach applicability.