Theory of three-pattern decomposition of global atmospheric circulation

This paper reviews the three-pattern decomposition of global atmospheric circulation(3P-DGAC)developed in recent years,including the decomposition model and the dynamical equations of global horizontal,meridional,and zonal circulations.Compared with the traditional two-dimensional(2D)circulation decomposition method,the 3P-DGAC can effectively decompose the actual vertical vorticity into two components that are caused by the horizontal circulation and convergent/divergent movement(associated with the meridional and zonal circulations).It also decomposes the vertical velocity into the components of the meridional vertical circulation and the zonal vertical circulation,thus providing a new method to study the dynamical influences of convergent/divergent motions on the evolution of actual vertical vorticity and an accurate description of local vertical circulations.The 3P-DGAC is a three-dimensional(3D)circulation decomposition method based on the main characteristics of the actual atmospheric movements.The horizontal,meridional,and zonal circulations after the 3P-DGAC are the global generalization of Rossby waves in the middle-high latitudes and Hadley and Walker circulations in low latitudes.Therefore,the three-pattern decomposition model and its dynamical equations provide novel theoretical tools for studying complex interactions between middle-high and low latitude circulations as well as the physical mechanisms of the abnormal evolution of large-scale atmospheric circulations under the background of global warming.

THE DOUBLE-LAYER STRUCTURE OF THE HADLEY CIRCULATION AND ITS INTERDECADAL EVOLUTION CHARACTERISTICS

Based on the three-pattern decomposition of global atmospheric circulation(TPDGAC), this study investigates the double-layer structure of the Hadley circulation(HC) and its interdecadal evolution characteristics by using monthly horizontal wind field from NCEP/NCAR reanalysis data from 1948—2011. The following major conclusions are drawn: First, the double-layer structure of the HC is an objective fact, and it constantly exists in April,May, June, October and November in the Southern Hemisphere. Second, the double-layer structure is more obvious in the Southern than in the Northern Hemisphere. Since the double-layer structure is sloped in the vertical direction, it should be taken into consideration when analyzing the variations of the strength and location of the center of the HC.Third, the strength of the double-layer structure of the HC in the Southern Hemisphere consistently exhibits decadal variations with a strong, weak and strong pattern in all five months(April, May, June, October, and November), with cycles of 20-30 a and 40-60 a. Fourth, the center of the HC(mean position of the double-layer structure) in the Southern Hemisphere consistently and remarkably shifts southward in all the five months. The net poleward shifts over the 64 years are 5.18°, 2.11°, 2.50°, 1.79° and 5.76° for the five respective months, with a mean shift of 3.47°.

基于多种混合模型的径流预测研究

变化环境下径流的波动不断加大,给径流的精准预报带来新的挑战。基于“分解-合成”策略的混合径流预报模型来提高预报精度是当前研究的热点之一。以往研究聚焦在单一的混合预报模型而忽视了它们的适用性研究。基于此,以渭河流域为例,在优选多元线性回归(MLR)、人工神经网络(ANN)和支持向量机(SVM)单一预报模型的基础上,分别基于经验模态分解(EMD)、集合经验模态分解(EEMD)和小波分解(WD)构建了多种混合模型,并融合了大气环流异常因子的信息。结果表明:(1)SVM模型预测精度高于ANN和MLR;(2)混合预测模型预测精度均高于单一模型,混合模型中WD-SVM的预测精度优于EMD-SVM和EEMD-SVM;(3)融合大气环流异常因子后WD-SVM模型预测精度最高,对极值预报精度的提高较为明显。

Hadley环流的双层结构及其年代际演变特征

利用1948—2011年NCEP/NCAR月平均水平风场资料,采用全球大气环流的三型分解等方法,研究了Hadley环流的双层结构及其年代际演变特征。结果表明：（1）Hadley环流双层结构客观存在,且4、5、6、10、11月南半球Hadley环流双层结构长期存在;（2）南半球Hadley环流双层结构较北半球更加明显,双层结构在垂直方向上呈现倾斜状,对Hadley环流强度和中心位置的变化进行研究时,应考虑双层结构的影响;（3）4、5、6、10、11月南半球Hadley环流强度随时间都呈现强弱强的年代际演变特征,各个月份环流强度都存在20-30年和40-60年时间尺度的周期;（4）4、5、6、10、11月南半球环流中心位置（双层结构平均位置）随时间一致地表现为向南移动,各个月份环流中心分别向南移动了5.18°、2.11°、2.50°、1.79°、5.76°,平均向南移动了3.47°。