北极海冰与大气环流耦合主模态的时空特征分析

THE MAIN MODES OF ARCTIC SEA ICE AND ATMOSPHERIC CIRCULATION

对1979—2009年月平均的CFSR(The Climate Forecast System Reanalysis)海冰密集度(SIC)和海平面气压(SLP)资料进行多变量经验正交函数分解(MV-EOF),得出耦合主模态,并通过对温度、位势高度和风场的回归分析,进一步探寻海冰与大气环流的关系,第一模态SLP的特征为北极涛动(AO),SIC呈离散的正负中心分布但大体为东西反位相,AO正位相时,喀拉海、拉普捷夫海、东西伯利亚海和鄂霍次克海海冰减少,巴芬湾、波弗特海、楚科奇海和白令海海冰增加。耦合第二模态的SLP呈偶极子分布,负、正异常中心在巴伦支海和波弗特海,SIC在巴伦支海、弗拉姆海峡、格陵兰海、拉布拉多海和白令海,鄂霍次克海地区有正异常,在喀拉海、拉普捷夫海、东西伯利亚海、楚科齐海和波弗特海为负异常。耦合第三模态SLP在冰岛地区存在负异常中心,在拉普捷夫海地区有正异常中心,SIC在巴伦支海北部、弗拉姆海峡、格陵兰海为负异常,其余地区全为正异常。对SLP和SIC分别进行EOF分解,并与耦合模态进行比较,SLP的EOF主模态的时空分布与耦合模态中SLP的时空分布十分相似,SIC的EOF模态的时空分布则与耦合模态中SIC的时空分布有较大差别,说明耦合模态对SIC的分布影响较大,即大气环流对海冰分布的影响为主要的过程,海冰对大尺度的大气环流的模态的影响不明显。

Thirty-one years of monthly mean CFSR（ The Climate Forecast System Reanalysis）SIC （sea ice concentration） and SLP （sea level pressure） data and multivariate experience orthogonal function （MV-EOF） analysis are used to document coupled main modes of sea ice and atmospheric circulation. Temperature, geopotential height and wind fields are used in regression analysis to further explore the relationship between sea ice and atmospheric circulation. The first mode of SLP resembles the structure of the Arctic Oscillation （AO） , and SIC shows an out-of-phase struc- ture between the American and Eurasian parts of the Arctic. When AO in its positive phase, there is a negative sea ice anomaly in the Kara Sea, Laptev Sea, East Siberian Sea and Sea of Okhotsk, and a positive anomaly in Baffin Bay, the Beaufort Sea, Chukchi Sea and Bering Sea. The second mode of SLP is a dipolar structure, with its anom- aly center overthe Barents and Beaufort seas. There is a positive SIC anomaly over the Barents Sea, Fram Strait, Greenland Sea, Labrador Sea, Bering Sea and Sea of Okhotsk, and negative anomaly in the Kara, Laptev, East Si- berian, Chukchi and Beaufort seas. In the third coupled mode, SLP has a negative anomaly center near Iceland and a positive center over the Laptev Sea. SIC shows a negative anomaly in the northern Barents Sea, Fram Strait and Greenland Sea, and positive anomalies elsewhere. To compare with the coupled mode, the SLP and SIC data are analyzed with the EOF. Spatial and temporal characteristics of SLP in the coupled and SLP modes are almost the same. There are some significant differences be- tween SIC spatial and temporal characteristics in the coupled and SIC modes. This indicates that the coupled mode has greater impact on SIC. In other words, the atmospheric circulation mode influence of the distribution of SIC and SIC itself havelittle effect on that circulation.