机构地区: 武汉大学水利水电学院水资源与水电工程科学国家重点实验室
出 处: 《水科学进展》 2012年第1期74-79,共6页
摘 要: 资料缺失是进行水质风险分析的薄弱环节。结合水质指标浓度分布的先验信息及已有水质资料,在分析水质指标相关关系基础上建立其随时间变化的浓度分布模型,应用Bayes理论及Gibbs抽样方法对模型涉及的大量参数和超参数的统计特征进行了同步估计。在假定缺失数据为随机变量的基础上,应用该方法得到的大量样本较好实现了对缺失数据的估计,并结合风险分析理论进一步量化了水质超标风险。通过实例验证了模型的合理性,为资料缺失地区的水质风险分析提供了新的思路和方法。 Incomplete data poses a challenge in river water quality risk analysis.Using the priori distribution of water quality indices and existing water quality data,a water quality model is established to dynamically simulate the concentrations of pollutants in the water.The statistical characteristics of the model parameters and super-parameters are estimated simultaneously using the Bayesian theory and the Gibbs sampling method.If the incomplete data are assumed to be missing completely at random,the proposed sampling method can provide large number of samples to the model parameter estimation.The substandard water risk analysis can thus be conducted by the model water quality.The result of a case study shows that the model is able to perform risk analysis of the river water quality with incomplete data,which provides a new approach to the risk analysis of water pollution.
领 域: [水利工程]