摘要
讨论了一个热带太平洋海气耦合集合预报系统集合预报的检验问题。该集合预报系统模式为一个中等复杂程度的耦合模式，其中大气部分为统计模式，海洋部分为动力模式。初始扰动利用集合Kalman滤波同化得到，模式误差扰动由一个一阶马尔可夫随机微分方程生成，预报集合样本为100个。利用1995～2005年的观测资料进行了确定性预报检验，包括相关系数和均方根误差。在概率预报检验方面，包括Talagrand概率分布、离散度、Brier评分（BS）、命中率以及空报率的统计检验，并且根据检验结果对预报系统进行了初步评价。确定性检验表明，集合样本均值的预报水平在热带中太平洋区域要高于热带东太平洋和沿岸区域。同时概率预报检验结果表明，集合预报系统有较高的概率预报技巧，对确定性预报是一个完善和补充。
The verification of the ensemble forecast results of a tropical Pacific air-sea coupled ensemble prediction system is discussed. The model of the ensemble prediction system is an intermediate coupled model. The atmospheric component is a statistical model, and the ocean component is a dynamical model. An ensemble Kalman filter （EnKF） data assimilation system is implemented to provide the initial ensemble. A first-order linear Markov stochastic model is used to represent model errors. The ensemble size is 100. The deterministic verification, including correlation and root mean square error, and the probabilistic verification, including Talagrand probability distribution, spread, Brier score （BS）, and hit rate and false alarm rate are applied to the ensemble prediction system. Then some evaluations are made according to the verification results. The verifications are covering the period of 1995 2005. The deterministic verifications show that the prediction skill of the ensemble mean over the central tropical Pacific is particularly higher than that of the ensemble mean in the eastern basin and coastal region. The probabilistic verifications show that the ensemble system has relative high probabilistic skill and it is a complement to the deterministic forecast.
出处
《气候与环境研究》
CSCD
北大核心
2007年第5期587-594,共8页
Climatic and Environmental Research
基金
国家自然科学基金资助项目40437017和40221503
中国科学院方向性项目KZCX2-YW-202
国家重点基础研究发展规划项目2006CB403600
关键词
集合预报
检验
评价
ensemble prediction, verification, evaluation
