河流藻类生长动态预测方法研究--以德国易北河为例  

Algal growth dynamic prediction in river——a case study of Elbe River

作　　者：赵晓东[1,2] 张宏建[1] 周洪亮[1] 

机构地区：[1]浙江大学控制科学与工程学系,杭州310007 [2]中国计量学院计量测试工程学院杭州,310018

出　　处：《水生态学杂志》2014年第1期13-21,共9页Journal of Hydroecology

基　　金：国家科技重大专项(2008ZX07420-004)

摘　　要：基于变量参数的建模思想,依据拉格朗日法提出了河流藻类生长单日定时预测机理,并采用L-M率定方法对模型参数进行动态率定。该模型以德国易北河为例进行验证。首先构建了河流藻类叶绿素a单日12时预测机理模型,对机理模型中的7个参数进行选择,取其中5个参数作为率定参数,其余2个参数作为先验参数。根据易北河2000年5月1日至8月1日期间的12时的叶绿素a浓度,对2组5参数组合分别进行率定,根据率定参数将机理模型沿时间轴外推至次日12时获得计算值作为预测结果。结果表明以率定最优(率定误差最小)为原则选取的5参数组合获得的预测结果最优。而通过改变率定数据长度获得的率定参数进行预测比较表明,采用连续7 d的近期历史数据获得预测结果最优。该方法还原藻类生长空间和时间异质性上优于先验参数预测方法,为采用机理模型进行河流短期藻类预测提供了一种新的解决方案。According to Lagrange method, a mechanism model based on the recent historical data was built to predict chlorophyll a concentration on a regular time next day. The optimum combination of calibrated parameters and the optimum calibration period were both determined by the L-M method. The results calculated by the model with the optimal calibration parameters were the predicted values. The mechanism model of algae growth in Elbe River based on the variable parameters was verified by comparing the predicted values for next day noontime and the real values observed in Elbe River between May and August in 2000. The minimum error of parameters calibration was used as the selection principles of parameters combinations. The predicted results obtained by selecting the reasonable parameters combinations （five parameters from seven parameters） showed much better than those obtained based on the prior parameters. The sampled data from different time span were also used to calibrate the five parameters. The predicted results under the calibration period of seven days showed better than others. The modeling method showed the advantage in the reconstruction of algal growth heterogeneity in both space and time compared to the priori parameter modeling method, and also provided a new solution for shore-term algal prediction in river when using mechanism model.

关 键 词：藻类 生长动态 预测模型 拉格朗日法 L-M算法 

分 类 号：Q145[生物学—生态学]