基于T106L26全球大气环流模式的夏季集合预报  

An ensemble forecast for summer with a global atmospheric general circulation model T106L26

作　　者：朱春子[1,2] 李清泉[2] 王兰宁[3] 王在志[2] 刘文泉[4] 

机构地区：[1]南京信息工程大学大气科学学院,江苏南京210044 [2]国家气候中心,北京100081 [3]北京师范大学全球变化与地球系统科学研究院,北京100875 [4]中国气象局,北京100081

出　　处：《大气科学学报》2013年第2期192-201,共10页Transactions of Atmospheric Sciences

基　　金：国家重点基础研究发展计划项目(2012CB955203);国家高技术研究发展计划项目(2010AA012404);公益性行业(气象)科研专项(GYHY(QX)201106022)

摘　　要：使用国家气候中心新一代大气环流模式BCC_AGCM2.0(T106L26)进行22a夏季(6—8月)、11个成员的集合回报试验;针对500hPa高度场、降水和气温的预测能力进行交叉检验,并计算其均方根误差。结果表明:模式对热带地区、海洋和欧亚大陆部分地区500hPa高度场的模拟较好;对我国长江中下游、华南大部分地区降水的模拟具有一定可信度;2m温度距平在我国北方大部分地区呈现正相关且相关系数通过90%的信度检验,在南方地区则有待改善。集合预报效果好于单样本预报。模式分辨率的提高在一定程度上有助于改进预报效果。The ensemble hindcast experiments of 11 members in summer （JJA） of 22 years were con- ducted by the BCC_AGCM2.0 （ T106L26 ） developed by National Climate Center. The capability of prediction in 500 hPa geopotential height, precipitation, and 2 m air temperature were tested by the cross validation and root mean square error methods. Results indicate that the model shows a good perform- ance of 500 hPa geopotential height in the tropical region, oceans and parts of Eurasia. The simulation of precipitation over the mid-lower reaches of Yangtze River and most parts of South China has consid- erable reliability. The correlation coefficients of the 2 m air temperature anomalies in most parts of northern China are positive and pass the local significance test at 90% confidence level. However,the results in southern China are needed to be improved. Meanwhile, the ensemble forecast is much better than the single-sample prediction. The forecasting results can be improved by using the higher resolution atmospheric general circulation model.

关 键 词：高分辨率大气环流模式 季节预测 交叉检验 集合预报 

分 类 号：P456.7[天文地球—大气科学及气象学]