城市河流水质监测指标间的相关性分析及预测

Correlation Analysis and Prediction of Water Quality Monitoring Indicators in Urban Rivers

为快速监测城市河流水质,有效追踪并控制污染源,选用某市4个水质监测站的19期监测数据,分析常规水质指标与核心水质指标间的相关性,揭示其水质状况和月度变化趋势,并使用Mantel检验和信息论模型选择方法确定与核心指标关联度最强的常规指标,建立最优回归预测模型。研究发现,不同监测站的相同核心指标与常规指标的关联程度各异;高锰酸盐指数与浊度、总氮与溶解氧、氨氮与电导率、浊度与氯化物总体均呈正相关,且具有小幅波动;五日生化需氧量和总磷与常规指标间无显著相关。分析表明,在全面分析各监测站大量有效数据的基础上,构建相对可靠的总体回归预测模型,所得水质监测分析结果可为制定有效的水质污染防治措施提供决策依据。

To quickly monitor water quality in urban rivers and effectively track and control pollution sources,19 groups of monitoring data from four water quality monitoring stations are selected in a city to analyze the correlation between conventional water quality indicators and core water quality indicators,and the water quality and the monthly changing trend are examined.Mantel tests and selection method of information theory model are used to determine the conventional indicators with the strongest correlation with core indicators and establish the optimal regression prediction model.It is found that the correlation between the same core and conventional indicators varies from monitoring station to station.Generally,they are in positive correlation with slight fluctuations,i.e.permanganate with turbidity,total nitrogen with dissolved oxygen,ammonia nitrogen with conductivity,and turbidity with chloride.No significant correlation is found between five-day biochemical oxygen demand and total phosphorus with conventional indicators.The analytical results show that a relatively reliable overall regression prediction model can be constructed based on a comprehensive analysis of large amount of effective data from various monitoring stations,and the analysis of water quality monitoring data can provide a basis for decision making of effective water pollution prevention and control measures.