基于Box—Jenkins方法的黄河水质时间序列分析与预测

采用基于Box-Jenkins方法的时间序列分析技术，对黄河上游甘肃兰州段、中游吴堡和下游山东利津段的水质进行了趋势分析和预测。选取对水质产生影响较大的两个污染因子化学需氧量（CODMn）和溶解氧（D0O1994—2003连续10a的月平均水质监测数据，借助Matlab和SAS统计软件，建立了ARIMA模型和乘积季节时间序列模型，并分析了这两个污染因子随时间推移的变化规律。结果表明：ARIMA模型和乘积季节模型能够用于短期水质预测，并且预测效果较好。黄河流域从上游到下游水质总体状况呈逐渐下降趋势，上游水质一般为Ⅱ和Ⅲ类，而中游和下游水质基本为Ⅳ、Ⅴ和超Ⅴ类。

Time Series Analysis and Forecast Model for Water Quality of Yellow River Based on Box-Jenkins Method

Water pollution has deteriorated the water quantity available.Based on the monitor data of water quality during 1994-2003, the present paper analyzed andforecasted the water quality in Lanzhou, Wubao and Lijin hydrologic （al） station of Yellow River by applying the time series analysis technology based on Box-Jenkins method. The daily monitoring data of dissolved oxygen （DO） and chemical oxygen demand（CODMn） were selected for investigating the spatial and temporal variability and for providing important water quality information on pollution problems in the Yellow River region. With the Matlab and SAS software, using 1994 to 2003 water quality monthly data, this paper created a product seasonal ARIMA time series models and analysis of variation which were dissolved oxygen （DO） and chemical oxygen demand （CODMn） over time. The results showed that ARIMA model and muhiplicative seasonal model were effective in short term prediction of water quality and the prediction accuracies were well. The analysis of the modeling results showed that the water quality of Yellow River belonged to class Ⅱ and Ⅲ in Ningxia and Lanzhou section, and class Ⅳ and Ⅴ in the Wubao to Lijin section.