| 摘要: |
| 【目的】研究洪水预报的多模型组合预报模式，丰富洪水预报的方式方法。【方法】将最优非负可变加权系数的组合预测方法应用于洪水预报中，分别建立最优加权组合预测模型和最优非负可变加权系数组合预测模型，以新安江模型、萨克拉门托模型、水箱模型和陕北模型为例进行组合预报，并用东洋河1982年8月的一场洪水实测流量值对2种组合模型进行了实例验证和比较。【结果】最优非负可变加权系数组合预测模型的3种误差指标均方误差（MSE）、平均绝对误差（MAE）和平均绝对百分比误差（MAPE）分别为18.394 0，51.317 2和0.131 4，明显小于最优加权组合预测模型的28.110 7，108.816 7和0.268 9，且前者的确定性系数（DC）为0.961 3，明显大于后者的0.909 5，表明应用最优非负可变加权系数组合预测模型的预测结果优于定权重的最优加权组合预测模型的预测结果。【结论】 考虑变权重的最优组合预测能进一步提高洪水预测的精度。 |
| 关键词: 洪水预报 多模型 组合预测 东洋河 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目（51109175）；中国博士后科学基金项目（20110490170）；水利部公益性行业科研专项（201001011）；陕西省科技攻关计划项目（2012K06-20）；陕西省教育厅科研计划项目（12JK0472） |
|
| Application of variable weighting combination forecasting model to flood prediction |
|
|
| Abstract: |
| 【Objective】The multi-model combination forecast of flood was studied to rich the methods of flood prediction.【Method】In this paper,the optimal nonnegative variable weighting combination forecasting method was applied to the flood prediction.The optimal weighting combination forecasting model and optimal nonnegative variable weighting combination forecasting model were established respectively,and applied t to the Dongyang River flood prediction.Taking Xin’anjiang model,Sacramento model,Tank model and Shaanbei model as example,the flood prediction of Dongyang River in 1982 was conducted.Then,the two combination forecasting models were validated and their accuracies were compared.【Result】Three error indexes:mean square error (MSE),mean absolute error (MAE) and mean absolute percentage error (MAPE) of the optimal nonnegative variable weighting combination forecasting model were 18.394 0,51.317 2 and 0.131 4,respectively,which were less than that of optimal weighting combination forecast model (28.110 7,108.816 7 and 0.268 9).And deterministic coefficient (DC) of the former model was 0.961 3,which was significantly greater than the latter (0.909 5).The results of the study showed that the accuracy of the optimal nonnegative variable weighting combination forecasting model was better than that of the optimal weighting combination forecasting.【Conclusion】The combination forecasting considering weighting variable can further improve the precision of flood prediction. |
| Key words: flood forecasting multi-model combination forecasting Dongyang River |
