论文部分内容阅读
基于三门峡、郑州、卢氏、孟津、济源和孟州站6～8月累积雨量资料,采用灰关联模式识别与预测方法进行了中长期雨量预测,预报验证期为5年。结果表明,6个站点的预报合格率均大于等于60%,总体上看丰水年的预报合格率大于枯水年,其中卢氏站的预报合格率最好,达到100%;三门峡站2006年、孟津站2002年和2006年预报相对误差较大,原因在于其实测值在各自站点从小到大排序中排位靠前,可见灰关联模式识别与预测方法在预测极小值时效果较差。
Based on the accumulated rainfall data from June to August in Sanmenxia, Zhengzhou, Lhushi, Mengjin, Jiyuan and Mengzhou stations, the mid-long term rainfall forecast was carried out by using the gray relational pattern recognition and prediction method. The forecast verification period was 5 years. The results show that all six stations have a pre-qualification rate of 60% or above. Overall, the pre-qualification rate in wet years is greater than that in low-flow years, of which Lushi Station has the best forecasting rate of 100% and Sanmenxia Station 2006 , Mengjin station in 2002 and 2006, the relative error is larger, because the actual measured values in their ranks from the smallest to the highest ranking, shows that the gray correlation pattern recognition and prediction methods in the prediction of the minimum effect is poor.