一种改进的合成流量法及其在水文预报中的应用

合成流量法对下游站流量作出预报的关键是确定各上游站流量到达下游站的时间。在实际中常采用平均传播时间,然而不同水情下真实传播时间与平均传播时间存在一定的差距,影响预报精度。对此,提出时间窗口概念,以平均传播时间为中心,向前、后各开一个时间窗口,用上游各站时间窗口内流量的线性组合表示其到达下游站的流量,以提高模型对不同水情的适应能力和预报精度。以三峡水库2009～2015年流量数据率定参数,以2016～2018年流量数据作为预报对象,试验结果表明时间窗口法能够显著改善预报精度。

An Improved Resultant Flow Method and Its Application in Hydrological Forecasting

Resultant flow method is an important method for short-term flood forecasting, whose key is to determine the time when the flows of the upstream stations reach the downstream stations. In practice, the average propagation time is usually adopted, but there is a certain difference between the real propagation time and the average propagation time under different water conditions, which affects the prediction accuracy. This paper proposes the concept of time window, which takes the average propagation time as the center, opens a window of a time period before and after, and represents the flow from the upstream station to the downstream station with the linear combination of the flow in the upstream station time window, so as to improve the adaptability of the model to different water conditions and forecast accuracy. This paper takes the flow data in the Three Gorges Reservoir area from 2009 to 2015 to calibrate parameters, and the flow data during 2016-2018 as the prediction object. The experimental results show that the time window method can significantly improve the accuracy of flow prediction.