将贝叶斯理论和模糊集理论引入水质评价领域,分别描述了水质评价过程中模型结构和参数的不确定性,建立了基于三角模糊数的贝叶斯模糊综合水质评价模型.对Ⅰ、Ⅱ、Ⅲ、Ⅳ、Ⅴ类水分别赋值,并根据监测点水质对各类别的后验概率计算水质的综合得分进而确定水质类别.同时,选取了TP、NH+4-N、COD、DO、As及粪大肠菌群为水质评价因子,将建立的模型应用于2010年洞庭湖水质评价中.结果表明,小河嘴、横岭湖、万子湖、目平湖、洞庭湖出口水质达到Ⅱ类水标准;南嘴、岳阳楼水质为Ⅲ类水;东洞庭湖和扁山水质介于Ⅱ类水和Ⅲ类水之间,对各等级的隶属度分别为0.4983Ⅲ+0.5017Ⅱ、0.7962Ⅲ+0.2038Ⅱ.各监测点位中,仅南嘴水质未达到相应的水质标准.较均值模型、三角模糊数模型和传统贝叶斯模型而言,基于三角模糊数的贝叶斯水质评价模型对不确定性的表达更为全面、更符合实际.
An integrated fuzzy-Bayesian water quality assessment model based on triangular fuzzy numbers was developed by introducing Bayesian Theory and Fuzzy Set Theory. The new model could express uncertainties from model structure and parameter uncertainty. The water quality grade at each sampling site was determined according to their posterior probabilities and assignments. TP, NH4 -N, COD, DO, As and fecal coliforms were selected as parameters. The new model was applied to evaluate the water quality of Dongting Lake in 2010. The results showed that the water quality of Xiaohe Zui, Hengling Lake, Wanzi Lake, Muping Lake and the outlet of Dongting Lake attained water quality standards Grade Ⅱ , and that the water quality of Nan Zui and Yueyang Tower reached water quality standards Grade Ⅲ. The water quality of East Dongting Lake and Biansban were between Grade Ⅱ andGrade m with the grade value of0.4983Ⅲ+0.5017Ⅱ、0.7962Ⅲ+0.2038Ⅱ, respectively. Compared with the current water quality standard, Nan Zui failedto reach the corresponding water quality standard. The water quality assessment model developed in this paper is more comprehensive and reasonable in comparison with traditional models.