论文部分内容阅读
详细介绍了一个新的大样本集合预报系统.为了减小ENSO(厄尔尼诺-南方涛动)预报中的预报不确定性,该集合预报系统首先基于一个中等复杂程度的耦合模式,利用集合卡尔曼滤波资料同化方法同化有效的海洋观测资料为集合预报系统提供集合初始场;同时,一个发展的用于12个月预报的一阶线性马尔可夫(Markov)随机误差模式被嵌套到集合预报系统中来模拟模式不确定性.基于1992年11月～2008年10月100个样本的集合回报试验,从确定性预报技巧和概率预报技巧2个方面对集合预报系统的预报水平进行了检验.该集合预报方法能够很有效地将传统的确定性预报扩展到概率预报领域,且检验结果表明,预报样本均值的预报水平要优于单一的确定性预报.对于概率预报而言,集合预报样本能够很好地跟随观测的变化,并且能够提供单纯确定性预报所不能够提供的额外信息.
A new large sample ensemble forecasting system is introduced in detail.In order to reduce the forecast uncertainty in the prediction of ENSO (El Ni --Southern Oscillation), the ensemble forecasting system first based on a coupling model of medium complexity, using ensemble Kalman filter Data assimilation method assimilates the available ocean observational data to provide a set initial field for ensemble forecasting system. Meanwhile, a developed first-order linear Markov random error model for 12-month forecast is nested into ensemble forecasting system To simulate the model uncertainty.According to the set return test of 100 samples from November 1992 to October 2008, the prediction level of the integrated forecasting system was tested from two aspects: deterministic forecasting skill and probability forecasting skill, The forecast method can effectively extend the traditional deterministic forecasting to the field of probabilistic forecasting, and the test results show that the forecasting sample mean is better than the single deterministic forecasting. For probabilistic forecasting, the ensemble forecasting sample can be very good Follow the changes in the observations and provide additional information that can not be provided by purely deterministic forecasts.