ARIMA模型在苏州河总磷预测中的应用

Application of ARIMA model in forecasting total phosphorus of Suzhou Creek

目前苏州河面临潜在的富营养化危机，氮、磷含量较高的污染源是富营养化形成的直接诱因。选择总磷指标为研究对象，介绍运用时间序列分析法对总磷进行ARIMA建模预测，确定ARIMA（11，0，0）即AR（11）为最终模型，用1986—2003年数据对2004和2005年进行预测，结果得出苏州河总磷在未来2年呈波动下降趋势，最终在0．4—0．6mg／L之间上下波动，但仍高于地面水V类标准。模型适用于苏州河总磷的短期预测，完善苏州河富营养化预测管理系统。

The Suzhou Creek faces the potential risk of eutrophication recently. The pollution effluent with highly-concentrated nitrogen and phosphorus is the main reason. Total phosphorus was selected as the key factor and introduced to the ARIMA model to set up a method on forecasting total phosphorus in the creek. Eighteen years data of total phosphorus, 1986 - 2003, from Wuninglu section was used to get parameters of the model and the ARIMA（ 11,0,0） model, AR（ 11 ） was confirmed as the final model to predict the total phosphorus in the creek. Forecasting result shows that the concentration of total phosphorus will be decreased and have a fluctuation within a narrow range in future,finally fluctuate between 0.4 mg/L and 0.6 mg/L. The value is also above sur- facewater standard V. This model can be applied on short time forecast with total phosphorus in Suzhou Creek and improve management system of eutrophication.