中文摘要：

针对通常水文预报过程参数的不确定性问题，利用贝叶斯理论，结合自适应采样的马尔可夫链蒙特卡罗方法来研究Nash模型参数的不确定性，并进行概率洪水预报。实例研究表明，该方法能充分利用已知的后验信息获取Nash模型参数的不确定性，得到其后验分布。根据获得的参数后验分布可实现概率洪水预报，同时给出各时刻洪水流量的均值和方差的预报值，为估计各种防洪决策的风险提供了依据。

英文摘要：

In this study, the uncertainty of natural hydrological prediction processes was modeled using the Bayesian theory and Markov Chain Monte Carlo based Adaptive Metropolis method. Then the probabilistic flood forecasting was made. The case study shows that the method proposed can provide the posterior distribution of parameters of Nash model based on posterior information of parameters. Using obtained parameter posterior distributions, a probabilistic flood forecast can be made, and the predictions of mean and variance of flood discharge can also be obtained, which can be used to estimate the risk of flood control decision.