针对水文模型参数不确定性分析常用方法 收敛速度缓慢，容易陷入参数空间局部最优区域等 问题，提出了PAM (parallel adaptive metropolis) 算法；对三水源新安江模型参数不确定性进行分析 研究。实例研究表明显著提高了计算速度和求解质 量，参数后验分布结果为区间预报提供了条件。

Markov Chain Monte Carlo (MCMC) methods,  which are popular for estimating parameters uncertainty of hydrologic models,  generally converge slowly,  and are easy to get stuck in a local optimized region in the parametric space during uncertainty assessment of hydrologic model parameters.  In this paper the Parallel Adaptive Metropolis (PAM) algorithm is presented to access the parameters uncertainty of hydrologic models.  The PAM algorithm provides an adaptive MCMC sampler to estimate the posterior probability distribution of parameters under Bayesian framework.  The performance of the PAM algorithm is greatly improved in the manner of parallel computing.  The PAM algorithm is applied to assess the parameter uncertainty of Xinanjiang model using hydrologic data from Shuangpai Reservoir.  The case study demonstrates that there is considerable uncertainty about the Xinanjiang model parameters.  The hydrograph prediction uncertainty ranges associated with the posterior distribution of the parameters estimates can bracket the observed flows well,  but not large,  indicating that the method is feasible.

程春田（1965－），男，湖北孝感市人，大连理工大学教授，博士生导师，主要研究方向：数字防汛减灾与水电站及电网经济运行研究