基于二项分布检验法的水质达标评价方法研究

A statistical method with water quality compliance assessment based on binomial test

水质达标评价是流域水污染防治决策的科学基础,但水质变量的不确定性在传统的达标评价中常被忽略并由此产生决策风险.二项分布检验法可定量表征水质变量不确定性导致的决策风险,其结果更具鲁棒性.基于二项分布检验法,纳入风险偏好和利益权衡,提出了水质达标评价的决策框架.以海河流域为研究对象,对其2014年7月22日—2015年7月14日的周监测数据进行达标评价分析,对比了不同方法的决策风险,发现平均值法会导致很大的取伪错误概率,超标比例法会导致很大的弃真错误概率;二项分布检验法的2类错误概率均可控制在相对较低水平,决策风险最小,可为决策者提供较大的决策空间.案例研究表明,二项分布检验法在水质达标评价中具有适用性和灵活性.

Water quality compliance assessment is essential for development and implementation of basin-scale water pollution control plan. Traditional water quality compliance assessment usually neglects the uncertainties associated with water quality variables,resulting in significant risks in decision support. This risk can be quantitatively characterized using binomial test,hence a more reliable decision support can be achieved through incorporating this method. In this study,we proposed a binomial test based decision-making approach for water quality compliance assessment. This approach explicitly incorporates the risk preference of the decision maker and benefit tradeoffs among different parties. We applied this approach to a case study where the compliance of weekly CODMnconcentration in River Haihe during July 22 th 2014 to July 14 th 2015 is evaluated. Same evaluations have also been conducted using another two existing approaches for comparison purpose,including the mean value method and a method based on the exceedance percentage. Results indicate that the mean value method would lead to large TypeⅡerror,while the method based on the exceedance percentage would lead to large TypeⅠerror. In contrast,the proposed binomial test based method could provide results with relatively small Type I and Type II errors,which is more reliable for decision support. Results from the case study indicate that Binomial test based approach provides a robust alternative for water quality compliance assessment.