新安江模型中实时校正技术的比较研究

A Comparative Study on The Real-time Co-correction of Xin'anjiang Model

由于流域特性和降水分布的复杂性和多变性,用降水径流或河道汇流做出的流量序列预报有时误差很大,需要进行实时校正以提高预报精度。采用了实时校正量法、卡尔曼滤波法、反馈模拟实时校正法对增江流域的新安江模型预报结果进行了实时校正,并将这3种结果进行比较分析。结果表明,实时校正技术能在一定程度上提高模型的预报精度,在本文试验中,反馈模拟实时校正法取得了较好的效果。

Due to the complexity and variability of the watershed characteristics and the precipitation distribution, the flow sequence predicted by rainfall runoff and river confluence sometimes errors great, which needs real-time correction to improve forecast accuracy. In this paper, real-time corrected value method, Kalman filtering method and simulating feedback real-time correction method are applied for the real-time co-correction of the prediction results of Xiffanjiang Model in the Zengjiang basin, and the three result are compared and analysised. The results indicate that the real-time correction technology can improve the forecasting accuracy to a certain extent, and the simulating feedback real-time correction method obtained better result in this test.