以东江干流(珠江流域支流)河源、岭下和博罗3个测站水位-流量数据为例,运用贝叶斯方法拟合水位流量关系曲线中的幂律模型.以东江干流历年实测数据构建合理的先验分布为基础,结合似然函数,导出后验分布,并用马尔科夫链蒙特卡洛(MCMC)算法估计后验分布中的参数.结果表明:贝叶斯方法能够合理推断水位流量关系曲线中的幂律模型并结合MCMC算法进行参数估计,且能够提供拟合的水位流量关系曲线的95%置信区间;相比最大似然估计法,贝叶斯方法在曲线的外延性表现更好.
This paper presents a Bayesian approach for fitting the power-law rating curve model to a set of stage- discharge measurements of Heyuan, Lingxia and Boluo, respectively, which are located in the East River. After crea- ting a reasonable prior distribution with the previous stage-discharge measurements, the posterior distribution is de- rived with the likelihood function. Then an efficient MCMC algorithm for the parameters estimation is used. The stud- y results show that~ Bayesian approach is reasonable for fitting the power-law rating curve model and estimating the parameters with MCMC algorithm. Also, the 95 % confidence interval is created by Bayesian approach~ Bayesian ap- proach performs better on extension of rating curves than the maximum likelihood estimation.