时间序列分析法在太湖总磷含量动态模拟中的应用

Application of Time Series Analysis to Dynamics Simulation of Total Phosphorus Content in Taihu Lake

以太湖为研究对象,根据其营养水平进行分区,使得不同区域内磷的浓度不同,而同一区域磷的浓度大体一致。利用GIS技术对每个区域内未采样点的总磷浓度进行插值,取区域内所有栅格的平均值作为其最终浓度。应用时间序列分析法对太湖1998—2004年每月的总磷含量进行动态模拟,建立各湖区的预测模型,结果表明:中营养区Ⅰ和轻富营养区Ⅱ符合AR(1)模型,中富营养区Ⅲ和富营养区Ⅳ符合AR(2)模型,重富营养区Ⅴ符合ARMA(2,3)模型。预测2005年各湖区总磷含量,经前10个月实测数据验证,说明所建模型能真实反映太湖总磷的动态变化趋势。

The phosphorus is a key nutrient element for the growth of waterborne organism in the lake and an important factor to control the eutrophication. The studies on dynamics change of lake total phosphorus （TP） content may provide some scientific guides to the effective control of the breaking out of eutrophication. As an example, Taihu Lake was divided into five areas according to its nutrition level, which the phosphorus concentration is different between the different areas and nearly identical in the same ones. The TP concentration in areas without sampling sites was interpolated by the technology of geographic information system and the average phosphorus concentration of all the grids within the areas was taken as its final concentration. Monthly TP content of the lake was simulated dynamically by the method of time series analysis since 1998 to 2004 and models for TP content in every area were established. The results show that AR（1） model is fit for mesotrophic region Ⅰ and lightly eutrophic region Ⅱ ,AR（2） model for mesoeutrophic region Ⅲ and eutrophic region Ⅳ and ARMA（2,3） model for hypertrophic region Ⅴ. The models predicted monthly total phosphorus content in every nutrition area in 2005 and the results were accorded well with monitoring data from January to October this year,which showed the models could evaluate correctly the trend of TP content dynamics change in Taihu Lake.