摘要
在经验正交函数分解(EOF)分析的基础上,采用旋转经验正交展开(REOF)方法,对江西省81个气象站6月降水进行客观分区,发现前4个主成分的累积方差达到了76.6%,这4个旋转主分量的高值区(绝对值≥0.5)基本覆盖整个江西省地区,且第1、2、3和4旋转主分量的高值区分别位于赣北南部、赣南、赣中及赣北北部,这表明江西省6月的降水可以划分为赣北北部(Ⅳ区)、赣北南部(Ⅰ区)、赣中(Ⅲ区)及赣南(Ⅱ区)4个区域。分别对各区6月降水与上年9月—5月SST和OLR场进行相关分析,找到了对各区6月降水有预报指示意义的因子(前期SST或OLR),并利用多元逐步回归方法建立了各区6月降水的预报模型。利用这些预报模型对1980—2008年各区6月降水进行了模拟,各区模拟值与观测值的吻合较好,两者的相关系数分别为0.55(I区)、0.43(Ⅱ区)、0.58(Ⅲ区)和0.54(Ⅳ区),这表明了这些模型对各区6月的降水有较好的模拟能力。2009—2013年各区6月降水模拟效果检验结果也显示,4个区模拟效果检验5 a中有4 a符号与实况一致。2014年6月预测结果的检验也显示,Ⅰ区、Ⅲ区、Ⅳ区预测的降水距平百分率与实况非常接近,进一步证明了这些模型的模拟性能。
Based on the decomposition analyses of Empirical Orthogonal Function （EOF）, the Rotated Empirical Orthogonal Function （REOF） method is used to analyze the spatial distribution features of Jiangxi precipitation in June by using the 1959-2014 precipitation data of 81 stations. It is found that the accumulated variance of the first four principal components reaches 76.6%, the high value area （the absolute value is larger than 0.5） of this four rotated principal components covers almost the whole Jiangxi province, and the high value areas of the first, the second, the third and the fourth rotated principal components are located in south part of northern Jiangxi, southern Jiangxi, central Jiangxi and north part of northern Jiangxi respectively, which meant that the June precipitation in Jiangxi can be divided to four subareas. With the SST and OLR data, study is undertaken on the relationship between the Jiangxi rainfall in June and SST as well as OLR. It turned out that Jiangxi rainfall in June is significant correlated to the previous （last September-May） SST and OLR at a significant level. The multiple regression method is used to establish the June precipitation forecasting models, the June precipitations for each subarea during 1980-2008 are simulated with these models, results show that the simulated values are in good agreement with those observed values, with the correlation coefficients being 0.55 （the first subarea）, 0.43 （the second subarea）, 0.58（the third subarea） and 0.54（the forth subarea） respectively, which means that these models have good abilities to simulate the June precipitation in Jiangxi for each subarea. The results of validity check of June precipitation during 2009-2013 for each subarea show that the simulated results of all the four subareas are consistent with the actual ones in four of these five years. Forecast test results also show that the simulated results in the first, the third and the forth areas are consistent with the fact in June 2014. This suggests that these models have a good ability on forecasting June precipitation.
出处
《气象与减灾研究》
2015年第1期8-15,共8页
Meteorology and Disaster Reduction Research
基金
江西省科技厅基金项目(编号:20111BBG70031-1)
关键词
降水趋势
气候预测
REOF
SST
OLR
REOF
SST
OLR
climate prediction
