Copula-MCP洪水概率预报方法研究与应用

常用的洪水概率预报方法一般都采用了正态化变换与线性假设,可能会导致信息的丢失,也影响了其对洪水过程的适用性。鉴于此,将Copula函数与模型条件处理器MCP相结合,不需要正态-线性条件约束,直接推求以预报值为条件的流量分布函数,构建Copula-MCP的洪水概率预报模型。以淮河王家坝断面为例,在经验降雨径流模型API的确定性预报结果的基础上,采用Copula-MCP模型实现洪水概率预报。对1990-2010年共25场洪水的研究结果表明:Copula-MCP模型优于MCP模型的概率预报结果,Copula-MCP模型的期望值预报亦优于API模型结果。

Research and Application of Copula-MCP Model in Flood Probabilistic Forecasting

Normal transformation and linear hypothesis are commonly used in flood probability forecasting methods,which may lead to the loss of information and affect its applicability to flood process.In this paper,the Copula function is combined with the model condition processor(MCP).Without the normal-linear condition constraint,the flow distribution function with the forecast value as the condition is directly deduced to construct the Copula-MCP flood probability prediction model.Taking wangjiaba section of huaihe river as an example,this paper adopts the Copula-MCP model to realize the flood probability forecast based on the deterministic forecast results of the Antecedent Precipitation Index Model(API).The results of the study on 25 floods from 1990 to 2010 show that the Copula-MCP model is superior to that of the probability prediction results of the MCP model,and the expected value prediction of the Copula-MCP model is also superior to that of the API model results.