摘要
为更准确地预报洪水发生概率,针对传统雨量不确定性计算方法中相对误差估计不准确的问题,将独立同分布中心极限定理引入降雨不确定性计算中,推求一定区域内某次降水过程中面雨量测值的相对偏差、测量误差以及相对误差,实现降雨不确定性概率描述;降雨量概率分布计算与确定性水文预报模型耦合,最终实现考虑降雨不确定性的洪水概率预报,并以滩坑流域2014年间5场洪水过程为例对该方法进行了验证。结果表明,5场洪水预报的确定性系数均在0.89以上,洪峰误差均在9%以内,洪量误差均在7%以内,且预报区间覆盖率均在61%以上。说明结合改进雨量不确定性计算方法的洪水概率预报效果较好,预报精度和覆盖率高,具有一定工程实际意义。
In order to forecast the probability of flood occurring more accurately and solve the problem that relative estimation error in the traditional calculation method of rainfall uncertainty calculation is not accurate,the independent identically distributed central limit theorem was introduced into the calculation of rainfall uncertainty by which the relative deviation,measurement error and relative error of the measured rainfall in a certain area is ascertained and the probability of rainfall uncertainty is able to be described.Finally,probabilistic flood forecasting considering the rainfall uncertainty was obtained by coupling the probability of rainfall distribution with the deterministic hydrological forecasting model,and verified the method by 5 floods in 2004 in the Tankeng Basin.The results show that the certainty coefficients of 5 floods are above 0.89,peak errors are within 9%,volume errors are within 7%,and the forecast interval coverage rates are above 61%.Besides,the results show that the forecasting effectiveness of probabilistic flood forecasting combined with improved rainfall uncertainty calculation method is more accurate,forecast accuracy and coverage rate are higher.So,it has certain practical significance.
出处
《水电能源科学》
北大核心
2017年第10期68-71,67,共5页
Water Resources and Power
基金
国家自然科学基金项目(51279047)
关键词
不确定性计算
独立同分布
相对误差
洪水概率预报
uncertainty calculation
independent identically distribution
relative error
probabilistic flood forecasting