摘要
贝叶斯方法是一种特殊的洪水频率分析的方法。它把单站信息和地区信息有机地结合在一起去推求设计洪水。贝叶斯方法认为参数向量θ是具有密度函数 f″(θ)的随机向量。θ的数学期望即为贝叶斯点估计,由 f″(θ)加权的 f(×/θ)构成了 x 的贝叶斯分布。本文用贝叶斯方法对 Mundeni Rru 河上的 Nilobe 站进行洪水频率分析,分析的结果是令人满意的。
The Bayesian method is a special method in flood frequency analysis. It combines at site and regional information to infer probabilities of extreme hydrological events.It is different from traditional methods in flood freque- ncy analysis.It considers the parameter vector θ as the random vector with a distribution f″(θ).The expectation of θ provides Bayesian point esti- mates of the parameters.The distribution of f(x/θ) weighted by f″(θ) forms the Bayesian distribution of x. The pdf of prior is a key concept in the Bayesian approaches. Prior information can be “borrowed” from similar catchments using regression ana- lysis A case study of inference of flood quantiles for Mundeni river at Nilobe, Srilanka is presented to illustrate Bayesian methods.The results of analysis were satisfactory.
关键词
洪水
频率计算
贝叶斯方法
prior distribution
posterior distribution
bayesian point
estimates
bayesian distribution
