摘要
This paper refers to the CNOP-related algorithms and formulates the practical method and forecast techniques of extracting predictable components in a numerical model for predictable components on extended-range scales.Model variables are divided into predictable components and unpredictable chaotic components from the angle of model prediction error growth.The predictable components are defined as those with a slow error growth at a given range.A targeted numerical model for predictable components is established based on the operational dynamical extended-range forecast(DERF)model of the National Climate Center.At the same time,useful information in historical data are combined to find the fields for predictable components in the numerical model that are similar to those for the predictable components in historical data,reducing the variable dimensions in a similar judgment process and further correcting prediction errors of predictable components.Historical data is used to obtain the expected value and variance of the chaotic components through the ensemble forecast method.The numerical experiment results show that this method can effectively improve the forecast skill of the atmospheric circulation field in the 10–30 days extended-range numerical model and has good prospects for operational applications.
This paper refers to the CNOP-related algorithms and formulates the practical method and forecast techniques of extracting predictable components in a numerical model for predictable components on extended-range scales. Model variables are divided into predictable components and unpredictable chaotic components from the angle of model prediction error growth, The predictable components are defined as those with a slow error growth at a given range. A targeted numerical model for predictable components is established based on the operational dynamical extended-range forecast （DERF） model of the National Climate Center. At the same time, useful information in historical data are combined to find the fields for predictable components in the numerical model that are similar to those for the predictable components in historical data, reducing the variable dimensions in a similar judgment process and further correcting prediction errors of predictable components. Historical data is used to obtain the expected value and variance of the chaotic components through the ensemble forecast method. The numeri- cal experiment results show that this method can effectively improve the forecast skill of the atmospheric circulation field in the 10-30 days extended-range numerical model and has good prospects for operational applications.
基金
supported by the National Natural Science Foundation of China (Grant Nos. 40930952, 41105055)
Global Change Study of Major National Scientific Research Plan of China (Grant No. 2012CB955902)
Meteorological Special Project of China (Grant Nos. GYHY201106016, GYHY201106015)
