Interannual Variability of the Hadley Circulation Associated with Tropical Pacific SST Anomaly

GONG Xiaoqing, WANG Qi, and LIU Yulong

1)The Key Laboratory of Physical Oceanography,Ministry of Education,Ocean University of China,Qingdao266100,P. R. China

2)The Key Laboratory of Ocean-Atmospheric Interaction and Climate,Universities of Shandong Province,Ocean University of China,Qingdao266100,P. R. China

3)Meteorological Institute of Tianjin,Tianjin300074,P. R. China

4)National Marine Data and Information Service,Tianjin300171,P. R. China

Interannual Variability of the Hadley Circulation Associated with Tropical Pacific SST Anomaly

GONG Xiaoqing1),2),3), WANG Qi1),2),*, and LIU Yulong4)

1)The Key Laboratory of Physical Oceanography,Ministry of Education,Ocean University of China,Qingdao266100,P. R. China

2)The Key Laboratory of Ocean-Atmospheric Interaction and Climate,Universities of Shandong Province,Ocean University of China,Qingdao266100,P. R. China

3)Meteorological Institute of Tianjin,Tianjin300074,P. R. China

4)National Marine Data and Information Service,Tianjin300171,P. R. China

The seasonal and interannual variability of zonal mean Hadley circulation are analyzed, and the important effects of sea surface temperature (SST), especially the tropical Pacific SST, on the meridional circulation are discussed. Following results are obtained： 1) the Hadley circulation presents a single clockwise (anticlockwise) cross-equator circulation in the Northern (Southern) Hemisphere winter,while it is a double-ring-shaped circulation quasi-symmetric about the equator in spring and autumn. The annual mean state just indicates the residual of the Hadley cell in winter and summer. 2) The first mode of interannual anomalies shows a single cell crossing the equator like the climatology in winter and summer but with narrower width. The second mode shows a double ring-shaped cell quasi-symmetric about the equator which is similar to the Hadley cell in spring or autumn. 3) Vertical motion of the Hadley circulation is driven by sea surface temperature (SST) through latent and sensible heat in the tropics, and the interannual anomalies are mainly driven by the SST anomaly (SSTa) in the tropical Pacific. 4) The meridional gradient of SSTa is well consistent with the lower meridional wind of Hadley circulation in the interannual part. For the spatial distribution, the meridional gradient of SSTa in the Pacific plays a major role for the first two modes while the effects of the Indian Ocean and the Atlantic Ocean can be ignored.

Hadley circulation; mass stream function; SSTa; meridional wind; vertical motion

1 Introduction

As early as 1735, Hadley proposed the simple single cell meridional circulation in both hemispheres-rising in tropical regions and sinking in the pole, during his study of trades (Hadley, 1735). In 1856, Ferrel first proposed the three-meridional circulation concept-the Hadley circulation (HC) in low latitudes, the Ferrel circulation in mid-latitude, and the polar circulation in high-latitude; this finding was confirmed by Ye and Zhu (1958) and Lorenze (1967). More theories pointed out that the meridional circulation plays important roles in the meridional exchange of mass, momentum, angular momentum and other physical variables (Oort and Peixoto, 1983; Chang, 1995; Hurrell, 1996).

Subsequently, the Hadley circulation has been investigated by many researchers, and a lot of results about its spatial and temporal variability have been obtained. Studies in the 1970s pointed out that the Hadley circulation has a significant seasonal variability; the vertical movement in the tropics is dominated by the winterhemisphere HC (Oort and Rasmusson, 1970; Newell and Kidson, 1972). Dima and Wallace (2003) divided the HC into the equatorially symmetric part in summer and winter, and the equatorially antisymmetric mode in spring and autumn; both the two parts are essential. A lot of researches have been done for the interannual variability of the HC. Oort and Yienger (1996) drew that the strength of the HC is positively correlated with the SSTa in the equatorial eastern Pacific, and negatively correlated with the Walker circulation. Theoretical and modeling studies indicated that the tropical SSTa has an important influence on the interannual variability of the HC. Tropical SST heating can determine the location of ITCZ (ITCZ), which is the ascending branch of the HC (Lindzen and Nigam, 1987; Hou and Lindzen, 1992; Numaguti, 1995; Levine and Schneider, 2011). Longitudinal thermal contrast can also change the intensity and location of the HC (Zhang and Delworth, 2005; Broccoliet al., 2006; Luet al., 2008; Nguyenet al., 2013). Wang (2002a) first showed the detailed variations of the Walker and Hadley circulations during ENSO events in the Pacific. Zhou and Wang (2006) and Ma and Li (2007) discussed the boreal winter HC, and found two principal modes： equatorially asym-metric and quasi-symmetric modes. The time evolution of the two modes changed closely with tropical SSTa, especially with ENSO. Fenget al. (2011a, 2011b) conducted a similar study for boreal summer HC, and reached similar conclusions. Previous researches mainly focused on the HC changes in different seasons and the important effects of tropical heating on the vertical movement. Fenget al. (2013) investigated the principal modes of HC in boreal spring and discussed the important roles of SST over the Indo-Pacific Warm Pool. The impacts of two types of ENSO (canonical ENSO and ENSO Modoki) on boreal spring HC were discussed by Feng and Li (2013). Considering the continuity of the atmospheric circulation, this paper will discuss the HC throughout the total time series and the effects of the horizontal gradient of SST on the horizontal motion.

Meridional circulation is composed of two components, the meridional windvand vertical velocityω. Meridional circulation is generally represented in two ways： one is the vector field of the two components (Wang, 2002a, 2002b), the other is the mass stream function (MSF) (Palmen and Vuorela, 1963; Vuorela and Tuominen, 1964; Qinet al., 2006); it is difficult for the former to accurately quantify the circulation intensity, so MSF will be used in this paper to discuss the character of the Hadley circulation. There are mainly two functions for calculating the MSF： the iterative scheme proposed by Wu and Tibaldi (1987) and the overlay scheme proposed by Wang (1994). Guet al. (2004) compared the two schemes and found both can well characterize MSF but the latter is simpler because it only involves meridional wind component; therefore, the overlay scheme will be chosen in this paper.

2 Data and Methods

In this study, the monthly mean meridional and the vertical winds are taken from the NCEP/NCAR reanalysis, which has the 2.5°×2.5° longitude/latitude resolution and total layers of 17 (Kalnayet al., 1996). The SST data is from NOAA Reynolds dataset with a horizontal resolution of 2.0°×2.0° (Smith and Reynolds, 2004). All datasets are dated from January 1951 to December 2010, 720 months in total.

Here, the meridional circulation is zonal-mean-removed. The zonal mean continuity equation in the spherical coordinates can be calculated as follows：

Eq. (1) defines a two-dimensional mass stream functionψ(MSF), and we have：

The mass stream function can be obtained by integrating Eq. (2)：

Readers can refer to Wang (1994) and Qinet al.(2006) for specific numerical calculation processes.

3 Seasonal Variability of the Hadley Circulation

Fig.1 shows the annual mean meridional circulation, where the black line indicates the MSF (Unit： 109kg s-1), positive values indicate clockwise rotation, with the air flowing southward in the lower levels of the troposphere and northward in the upper levels, and negative values indicate anticlockwise rotation. The maximum value of the clockwise rotation center indicates the capacity of mass transport southward in lower levels and northward in the upper levels. The minimum value of the anticlockwise rotation center also indicates the mass transport capacity but with opposite direction. Shaded is the zonal mean vertical velocity (Unit： -10-4hPa s-1). The positive region corresponds to the rising motion, and negative region corresponds to the sinking motion.

Fig.1 Annual mean distributions of zonal mean vertical velocity (shaded, red for rising motion, blue for sinking motion, Unit： -10-4hPa s-1) and mass stream function (MSF, contour, positive value for clockwise cell, negative for anti-clockwise cell, Unit： 109kg s-1). Abscissa axis represents the latitude, the vertical axis represents the geopotential height, unit： hPa.

Four meridional cells can be found between 60°S to 60°N from the annual mean distribution： two Hadley cells (the SH and the NH) in the low latitudes, and two Ferrel cells (the SF and the NF) in the mid-latitudes. The Ferrel cells are located in the 30°-60° latitude band in each hemisphere, circulation center is located in 700 hPa geopotential height layer, the center intensity of the SF being stronger than that of the NF. As for the Hadley circulation, the SH crosses 30°S-10°N for about 40 latitudes, while the NH just crosses 10°N-30°N for about 20 latitudes,i.e., width of the SH is close to twice the NH, and the center intensity of SH is stronger than that of the NH, which isdifferent from the result calculated by Oort and Yienger (1996). The rising motion of the HC is located within 10°S-15°N; the main rising center is located near 10°N, corresponding to the annual mean ITCZ, and sinking center is near 30° in both hemispheres, corresponding to the zonal mean subtropical high. It is noteworthy that the subsidence center in southern hemisphere is significantly lower than that in the northern hemisphere, specific heights corresponding to 700 hPa and 400 hPa respectively in the southern and northern hemisphere.

Therefore, the SH is stronger than the NH, like the contrast between the SF and the NF in the annual mean meridional circulation. This comparison should be consistent with the ITCZ located near 10°N intuitively. As the Ferrel circulation shares the sinking branch with the Hadley circulation, and the Hadley circulation is far stronger than the Ferrel circulation, the following analysis will focus on the characteristics of the Hadley circulation.

Fig.2 shows the seasonal variation of the meridional circulation. The Hadley circulation has significant seasonal variations. In boreal winter (DJF) the NH is the strongest and dominates the tropical meridional circulation; these are main features of a cross-equatorial clockwise cell, which is located within 15°S-30°N, rising at 10°S and sinking at 20°N. Meanwhile, the SH is the weakest. In boreal summer (JJA), the situation is contrary. The SH is the strongest and displays an anticlockwise circulation across 30°S-20°N, with the center located a bit south of the equator at 500 hPa level, with the main rising and sinking centers respectively located at approximately 10°N and 20°S; the NH is simultaneously the weakest. Spring and autumn are transitional seasons. In boreal spring, the NH is gradually weakened and retreats northward, while the SH is gradually strengthening. It is contrary in boreal autumn, with the SH gradually weakening and retreating southward, and the NH gradually strengthening. It should be noted that, the NH moves southward more slowly in autumn, compared to SH moving northward in spring; but an obvious jump process occurs from November to December.

Fig.2 Seasonal variations of zonal mean vertical velocity (shaded) and mass stream function (contour) (same as Fig.1).

From Fig.2, it can also be noted that there are two rising centers at 10°N and a bit north to the equator in boreal winter and spring (from December to May)with the height of the center moves from the low-level (700 hPa) to the middle (500 hPa), while in boreal summer and autumn (from June to November), there is only one rising center located near 10°N, rising area is more concentrated, and in June there is a weak center at low levels which is subsequently strengthened except for the center at 400 hPa-500 hPa. Further, the NH is narrower than the SH, but extends higher; these characteristics can be explained by stronger ITCZ in the northern hemisphere.

From the above, in summer and winter, the monsoonlike meridional circulation is conducive to the exchanges of mass, momentum and angular momentum between the northern and southern hemispheres; in the transitional seasons spring and autumn, the meridional circulation is more like the original definition of the Hadley circulation, and this quasi-symmetric circulation about the equator is conducive to exchanges between tropical and extratropical regions.

Comparing Fig.1 and Fig.2, the annual mean result just describes the residual of Hadley circulation in winter and summer, because of the offset between the clockwise circulation in boreal winter and the anticlockwise circulation in boreal summer above the equator.

4 Interannual Variability of the Hadley Circulation

By removing the long-term mean part of the MSF to get the anomaly field, two most important modes of interannual anomalies can be found by EOF decomposition, the first mode (PC1) accounting for 39.6% and the second mode (PC2) accounting for 18.7% (Fig.3).

From Fig.3(a), positive (negative) phase of PC1 mainly draws a clockwise (anticlockwise) circulation located at 15°S-18°N, which is similar to the seasonal HC in the northern (southern) winter, but with narrower and higher center than the seasonal circulation (which is located at 300 hPa a bit north of the equator). Two weak circulation centers located at 20°S and 25°N, causing the rising region boundary of the lower one extending to 20°S and the sinking region boundary to 25°N in the positive phase. Therefore, comparing to Fig.2, in the northern (southern) winter, the positive (negative) phase of the first mode will strengthen the HC., The positive (negative) phase of the first mode will strengthen (weakened) the SH and weakened (strengthen) the NH in spring and autumn. The positive phase (Fig.3b) of PC2 is similar to the equator symmetric double-ring HC, located at 20°S- 18°N; the southern branch is wider and stronger. Two abnormal circulation centers are both located at 500 hPa. In addition, there is a weak anticlockwise circulation anomaly within 20°N-40°N, which is beneficial to the sinking of the HC in the lower latitudes 20°N-30°N. So the positive phase of the second mode would strengthen (weaken) the HC in northern (southern) winter and strengthen that in spring and autumn; contrast case would happen for the negative phase.

Fig.3 Spatial and temporal distribution of interannual variation of MSF (same as the contour in Fig.1) (a) spatial distribution for PC1, (b) spatial distribution for PC2, (c) temporal distribution for PC1 (solid line) and PC2 (dashed line).

Fig.4 Spectrum for the first two modes of interannual variability (dashed： 95% test line) (a) PC1, (b) PC2.

The time coefficient of both the two EOF modes has significant interannual variability. Power spectrum analysis shows that (Fig.4, dashed line is test line over 95%) the first mode has a significant period of 1.5a-3a, and the second mode has a significant period of 1.5a-4a.

5 Correlation Between Interannual Anomalous of the Hadley Circulation and SSTa

5.1 Distributions ofωaand SSTa

Figs.5 and 6 show the regression fields of vertical velocity anomaly (ω) and SSTa by the time distribution of the first two modes, and the Figs.(a), (b), (c), (d) are re-spectively forωat 300 hPa, 500 hPa, 700 hPa and 850 hPa, and (e) for the SSTa distribution.

Fig.5 Regression ofωand SSTa by PC1 time series of interannual variation, (a), (b), (c), (d) respectively forωaat 300 hPa, 500 hPa,700 hPa and 850 hPa (red for rising motion, blue for sinking motion, Unit： -10-4hPa s-1); (e)for SSTa (red for positive values, blue for negative values, Unit： ℃).

Fig.6 Same as Fig.5, but for PC2.

The positive phase of PC1 displays an abnormal circulation which presents the ascending region between the south of the equator and 20°S and the descending region within about 3°N-25°N (Fig.3a). From Fig.5, the abnormal rising motion at south of the equator mainly occurs in the middle east Pacific with the strongest at 300 hPa; rising motion also occurrs in the southwest Indian Ocean, but the intensity is weaker; the rising regions near the equator mainly appear in the eastern equatorial Pacific (500 hPa strongest) and the East Indian Ocean (300 hPa strongest). The sinking region in the Northern Hemisphere mainly occurrs in the west of 120°W over the North Pacific Ocean. The sinking centers at lower-level (850 hPa) and higher-level (300 hPa) are abnormally in the west (surrounded by red box). Sinking motion also appears on the Eastern Indian Ocean and the western Atlantic Ocean with weak strength.

SSTa distribution is more consistent with the vertical motion (Fig.5e)： the negative SSTa in the North Pacific corresponds to sinking motion in high-levels; the positive SSTa in the southern and eastern Pacific, as well as the eastern equatorial Pacific and the eastern Indian Ocean may explain the abnormal rising motion over them. The Atlantic contributes little to the HC in PC1 of interannual anomalies.

The positive phase of PC2 anomalous circulation shows abnormal rising from 7°S to 7°N, and abnormal sinking in other regions from 20°S to 30°N. Compared with Fig.6a, significant abnormal rising region occurrs over central and eastern equatorial Pacific Ocean; abnormal sinking regions are distributed in the subtropical area, mainly appearing over the east of the Philippines Sea near 10°N and over the North Pacific 160°E-140°W near 20°N, as well as over the eastern Indian Ocean, western Pacific warm pool and east of Australia in the southern Hemisphere. Another significant feature is the intensity of the center which is enhanced with height and the position sloping westward with height.

SSTa is also more consistent with the vertical motion in PC2 (Fig.6e)： the most prominent is the positive SSTa in equatorial eastern Pacific, driving the rising motion directly; there are weak negative SSTa at the South Pacific Convergence Zone (SPCZ) of the southern hemisphere. So, PC2 of the HC is the meridional circulation anomalies driven by El Nino-like SSTa. SSTa at Indian Ocean and the Atlantic Ocean are very weak for PC2.

5.2 Correlation Between Meridional Wind and Meridional Gradient of SSTa

In the Tropics, ocean heating on the atmosphere is an important factor in driving atmospheric circulation： First, warm ocean surface can drive atmospheric vertical motion directly through heating the atmosphere by latent and sensible heat; on the other hand, meridional gradient of SST may produce horizontal pressure gradient. For the meridional velocity anomaly of low-levelVadriven by heat directly, there is the relationshipVa∝? (SSTa)/?y, that is, the lower-level air moves from the area of lower SST to that of higher SST. Figs.7a and 7b are spatial distribution of the HC on interannual anomalies of PC1 and PC2, Figs.7c and 7d are the corresponding zonal mean meridional gradients of SSTa (?(SSTa)/?y) of regression distribution. The north wind between 10°S-15°N in the lower-levels showed from the positive phase of PC1 (Fig.7a) is well consistent with the negative ?(SSTa)/?y(Fig.7c). The positive phase of PC2 (Fig.7b) shows that in the lower-levels, meridional wind converges towards the equator in the belt of 15°S-15°N corresponding to the positive ?(SSTa)/?yin the south of the equator and negative ?(SSTa)/?yin the north of the equator. It can be confirmed that the relationshipVa∝?(SSTa)/?yis just suitable for the meridional wind driven by heat directly, butnot for the non-thermal directly driven meridional circulations. From the spatial distribution, ?(SSTa)/?yin the Pacific plays the major role for the first two modes, effects of the Indian Ocean and the Atlantic Ocean can be ignored (not shown).

Fig.7 MSF anomalies (a, b) and zonal mean meridional gradient of SSTa regression distribution (c, d) of interannual variation (a) (c) for PC1, (b) (d) for PC2.

6 Conclusions

The seasonal and interannual variability of the zonal mean Hadley circulation and their relationship with SST are discussed. The following conclusions are obtained：

1) Significant seasonal variation was found for the Hadley circulation. The Hadley circulation assumes a single clockwise (anticlockwise) circulation crossing the equator in the Northern (Southern) Hemisphere winter. In spring and autumn, however, it shows a double ring-shaped circulation quasi-symmetric about the equator, like the traditional Hadley circulation. The annual mean state of the Hadley cell just indicates the residual of winter and summer HC.

2) The first EOF mode of the interannual anomalies for Hadley circulation shows a single cross-equator cell like the long-term mean Hadley circulation in winter or summer, with narrower width and higher geopotentialheight center. The second mode exhibits double cells quasisymmetric about the equator which resemble the longterm mean Hadley circulation in spring and autumn.

3) Vertical motions can be driven by SSTa through latent and sensible heat. The main contributive regions to the positive phase of mode 1 are the areas with negative SSTa in the central Pacific north of the equator, the positive SSTa in the eastern Pacific south of the equator, the eastern equatorial Pacific, and in the eastern equatorial Indian Ocean. The main contributive region to the second mode is the area with significant SSTa in the central and eastern equatorial Pacific.

4) SSTa gradient can drive horizontal circulation. The analyses confirm that for the heat-driven Hadley circulation, there is a good relationship between the lower meridional windVaand the meridional gradient of SSTa,?(SSTa)/?y. For the spatial distribution, the gradient?(SSTa)/?yin the Pacific plays a major role, while the effects of the Indian Ocean and the Atlantic Ocean can be ignored.

Acknowledgement

This work was jointly supported by the National Basic Research Program (973 Program, Nos. 2012CB417402 and 2013CB956201).

Broccoli, A. J., Dahl, K. A., and Stouffer, R. J., 2006. Response of the ITCZ to Northern Hemisphere cooling.Geophysical Research Letters, 33, L010702, DOI： 10.1029/2005GL024546.

Chang, E. K. M., 1995. The influence of Hadley circulation intensity changes on extra tropical climate in an idealized model.Journal of Atmospheric Sciences, 52： 2006-2024.

Dima, I. M., and Wallace, J. M., 2003. On the seasonality of the Hadley cell.Journal of Atmospheric Sciences, 60： 1522-1527.

Feng, J., and Li, J. P., 2013. Contrasting impacts of two types of ENSO on the boreal spring Hadley circulation.Journal of Climate, 26： 4773-4789.

Feng, J., Li, J. P., and Xie, F., 2013. Long-term variation of the principal mode of boreal spring Hadley circulation linked to SST over the Indo-Pacific Warm Pool.Journal of Climate, 26：532-544.

Feng, R., Li, J. P., and Wang, J. C., 2011a. The principal modes of variability of the boreal summer Hadley circulation and their variations.Chinese Journal of Atmospheric Sciences, 35 (2)： 201-216 (in Chinese).

Feng, R., Li, J. P., and Wang, J. C., 2011b. Regime change of the boreal summer Hadley circulation and its connection with the tropical SST.Journal of Climate, 24： 3867-3877.

Gu, X. J., Wang, P. X., and Li, L. P., 2004. Comparison of two computational schemes of mass stream function of zonally mean meridional circulation.Journal of Nanjing Institute of Meteorology, 27 (1)： 11-19 (in Chinese).

Hadley, G., 1735. Concerning the cause of the general tradewinds.Philosophical Transactions: Royal Society, 39： 58-62.

Hou, A. Y., and Lindzen, R. S., 1992. The influence of concentrated heating on the Hadley circulation.Journal of Atmospheric Sciences, 49 (14)： 1233-1241.

Hurrell, J. W., 1996. Influence of variation in extratropical wintertime teleconnections on Northern Hemisphere Temperature.Geophysical Research Letters, 23： 665-668.

Kalnay, E., Kanamitsu, M., and Kistler, R., 1996. The NCEP/ NCAR 40-year reanalysis project.Bulletin of theAmerican Meteorological Society, 77： 437-472.

Levine, X. J., and Schneider, T., 2011. Response of the Hadley circulation to climate change in an aquaplanet GCM coupled to a simple representation of ocean heat transport.Journal of Atmospheric Sciences, 68： 769-783.

Lidezen, R. S., and Nigan, S., 1987. On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics.Journal of Atmospheric Sciences, 44 (17)：2418-2436.

Lorenz, E. N., 1967. The nature and theory of the general circulation of the atmosphere.WMO Publication, 218： 43-81.

Lu, J., Chen, G., and Frierson, D. M., 2008. Response of the zonal mean atmospheric circulation to El Ni?o versus global warming.Journal of Climate, 21 (22)： 5835-5851.

Ma, J., and Li, J. P., 2007. The reason for the strengthening of the boreal winter Hadley circulation and its connection with ENSO.Prog. Nat. Sci., 17 (11)： 1524-1531 (in Chinese).

Newell, R. E., Kidson, J. W., Vincent, D. G., and Boer, G. J., 1972.The General Circulation of the Tropical Atmosphere and Interactions with Extra Tropical Latitudes. The MTI Press, Vol 1, 258pp.

Nguyen, H., Evans, A., Lucas, C., Smith, I., and Timbal, B., 2013. The Hadley circulation in reanalyses： Climatology, variability, and change.Journal of Climate, 26： 3357-3376.

Numaguti, A., 1995. Dynamics and energy-balance of the Hadley circulation and the tropical precipitation zones. Part II：Sensitivity to meridional SST distribution.Journal of Atmospheric Sciences, 52 (8)： 1128-1141.

Oort, A. H., and Peixoto, J. P., 1983. Global angular momentum and energy balance requirements from observations.Advance in Geophysics, 25： 355-490.

Oort, A. H., and Rasmusson, E. M., 1970. On the annual variation of the monthly mean meridional circulation.Monthly Weather Review, 98： 423-442.

Oort, A. H., and Yienger, J. J., 1996. Observed interannual variability in the Hadley circulation and its connection to ENSO.Journal of Climate, 9： 2751-2767.

Palmen, E., and Vuorela, L., 1963. On the mean meridional circulations in the northern hemisphere during the winter season.Quarterly Journal of the Royal Meteorological Society, 89： 131-138.

Qin, Y. J., Wang, P. X., Guan, Z. Y., and Yue, Y., 2006. Comparison of the Hadley cells calculated from two reanalysis data sets.Chinese Science Bulletin, 51 (14)： 1741-1746.

Smith, T. M., and Reynolds, R. W., 2004. Improved extended reconstruction of SST (1854-1997).Journal of Climate, 17 (12)： 2466-2477.

Vuorela, L., and Tuominen, I., 1964. On the mean zonal and meridional circulations and the flux of moisture in the northern hemisphere during the summer season.Pure and Applied Geophysics, 57： 167-180.

Wang, C., 2002a. Atmosphere circulation cells associated with the El Nino Southern Oscillation.Journal of Climate, 15： 399-419.

Wang, C., 2002b. Atlantic climate variability and its associated atmosphere circulation cells.Journal of Climate, 15： 1516-1536.

Wang, P. X., 1994. Diagnosis of mean meridional circulation of the model atmosphere in the GCM with low resolution vertically.Journal of Nanjing Institute of Meteorology, 17 (2)：199-204 (in Chinese).

Wu, G. X., and Tibaldi, S., 1987. Effect of mechanical forcing on the meridional circulation and transfer properties of the atmosphere.Advances in Atmosphere Sciences, 4 (1)： 24-42.

Ye, D. Z., and Zhu, B. Z., 1958.Some Fundamental Problems of the General Circulation for Atmosphere. Science Press, 159pp (in Chinese).

Zhang, R., and Delworth, T., 2005. Simulated tropical response to a substantial weakening of the Atlantic thermohaline circulation.Journal of Climate, 18： 1853-1860.

Zhou, B. T., and Wang, H. J., 2006. Interannual and interdecadal variations of the Hadley Circulation and its connection with tropical sea surface temperature.Chinese Journal of Geophysics, 49 (5)： 1271-1278 (in Chinese).

(Edited by Xie Jun)

(Received August 30, 2013; revised October 21, 2013; accepted April 7, 2015)

? Ocean University of China, Science Press and Spring-Verlag Berlin Heidelberg 2015

* Corresponding author. E-mail： wangqi@ouc.edu.cn