基于面雨量预报的长江上游洪水分级预估及其应用

摘要

基于面雨量预报和线性回归方法,建立长江上游洪水分级预估模型。在此基础上,挑选1971—2009年6—8月长江上游5个流域面雨量满足暴雨条件的203次过程,分别用一元多次方程、多元一次方程计算由暴雨产生的洪峰流量,并与实况流量增量进行对比,同时对该模型2010年汛期三次洪峰过程实时预报结果进行检验。结果表明：该预估模型对长江上游洪水分级具有一定的预估能力,其预报的宜昌洪峰流量以偏大为主,偏大（0.1～0.6）×104m3.s-1不等;同时,对长江上游洪水分级预估的量级把握较好,72 h内量级预报准确率达100%;洪峰流量逐日滚动预报误差检验结果显示,除一天的预报相对误差超过20%外,其余预报与实况的相对误差均小于20%;长江流域水文预报台7d面雨量定量预报产品的使用,可使宜昌洪峰预报提前5天。%A flood classified projection model in the upper reaches of the Yangtze River was established with area rainfall forecast and linear regression method.We chose 203 processes whose area rainfall meets the heavy rain condition from five river basins in upper reaches of the Yangtze River during the period from June to August from 1971 to 2009.We also calculated the flood peak flow resulted from the heavy rain with the relevant regression equation and linear equation and contrasted the results with the observed flow increment.The real-time results made using this model were tested comparing with the three flood peak processes of flood season in 2010.The results show that the model has certain ability to predict flood classification in the upper reaches of the Yangtze River.The predicted flood flow result in Yichang was too large,which is（0.1～0.6）×104 m3.s-1.It has a better ability to predict the magnitude used for flood classification.The accuracy of predicted magnitude in 72 h reached 100%.The flood grades forecast of the upper reaches of the Yangtze River was realized with the model.The test result for 2010 showed that the qualified rate was all more than 80% except one day.By using the model the flood grades forecasted is 5 days ahead of the flood peak.