1. Introduction
Cross-equatorial flows (CEFs) play a key role in mass, moisture, and energy transport between the hemispheres (Tao et al. 1962; Findlater 1966; Wang and Li 1982; Wang and Xue 2003). A total of five branches of CEFs are seen over the tropical Indian–western Pacific Ocean in boreal summer, including the Somali CEF, Bay of Bengal (BOB) CEF, South China Sea (SCS) CEF, Celebes Sea CEF, and the New Guinea CEF (Li and Lou 1987; Shi et al. 2001; Peng and Jiang 2003; Shi et al. 2007; Li and Hu 2011). The latter three are located in the north of Australia and are referred to in combination as the Australian CEF. All the CEFs serve as an important member of the Asia–Pacific summer monsoons, including the Indian summer monsoon (ISM), the East Asian summer monsoon (EASM), or the western North Pacific summer monsoon (WNPSM) (Zhu et al. 1986; Tao and Chen 1987; Wang and Ho 2002).
Numerous investigators have studied the importance of individual CEFs for the EASM. In comparison to the ISM, for which the Somali CEF plays the dominant role for moisture transport (Saha and Bavadekar 1973; Cadet and Reverdin 1981), the amount of rainfall (Ramesh Kumar et al. 1999; Halpern and Woiceshyn 2001), and even the intraseasonal active–break phase of the monsoon (Joseph and Sijikumar 2004; Cadet and Greco 1987), these CEFs play various roles for the EASM (Tao and Chen 1987). The Somali CEF is not dominant and plays just a minor role in determining the EASM rainfall (Simmonds et al. 1999). Instead, the Australian CEF is more important. The stronger Australian CEF favors weakening of the EASM along with an eastward and northward movement of the western Pacific subtropical high (WPSH), a key member of the EASM system (Liu et al. 2009). Besides, the SCS CEF is positively correlated with the tropical component of the EASM, but is oppositely correlated with the ISM (Mao et al. 1990). The importance of the CEFs for various regional rainfalls in China was also studied. Han (2002) suggested that 1) the intensified Somali CEF favors above-normal rainfall in northern China but deficiency in the Yangtze–Huaihe River basin, 2) the BOB CEF is negatively related with rainfall in the middle and lower reaches of the Yangtze River, and 3) the Australian CEF is negatively correlated with rainfall in the region in between the middle reaches of the Yellow River and the Yangtze River but positively with rainfall in southern China. A similar result was obtained by Zhu (2012).
It is not of much meaning to understand the individual role of the CEFs for predicting the EASM, because the CEFs are not independent on but interact with each other (Zhang 2001; Wang and Yang 2008). Cong et al. (2007) found a strong negative correlation between the Somali and SCS CEFs, and the negative correlation varies at the decadal time scale. Recently, Zhu (2012) displayed two interdecadal shifts in both the Somali and Australian CEF, and they occurred almost simultaneously: one around the late 1970s and the other around the late 1990s. This suggests coherence between them on a decadal time scale.
The present study aims to investigate the linkage between the CEFs on the interannual time scale, as well as its connection to the EASM. Details about the dataset used are described in section 2. The results about the linkage between the CEFs are given in section 3. Because a seesaw connection is found between the Somali and Australian CEF, section 4 discusses the connection of the seesaw with the EASM. Considering the potential influence of SST, section 5 analyzes the roles of El Niño–Southern Oscillation (ENSO) and the tropical Indian Ocean (IO) SST dipole (IOD) in shaping the seesaw connection. A summary and discussion are given in the last section.
2. Datasets and methods
Various datasets for atmospheric circulation, convection, rainfall, and sea surface temperature (SST) are utilized. Two atmospheric circulation reanalysis datasets are used: the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) monthly-mean reanalysis (Kalnay et al. 1996) and the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) (Uppala et al. 2005). The National Oceanic and Atmospheric Administration (NOAA) interpolated outgoing longwave radiation (OLR) data are used for understanding convection activity (Gruber and Krueger 1984). Two monthly-mean precipitation datasets, one from the Global Precipitation Climatology Project (GPCP, version 2) for 1979–2013 (Adler et al. 2003) and the other from the mainland China 160-station precipitation for 1951–2011 provided by China Meteorological Administration, are used to analyze the CEFs connection with East Asian rainfall. The GPCP precipitation product is used due to its better agreement with that recorded in China in comparison with other global rainfall datasets (Ma et al. 2009). The distribution of the 160 stations in China is displayed together with the composite rainfall anomaly in the next section (see Fig. 5e below). The NCEP–NCAR reanalysis, ERA-40, OLR, and GPCP data are on a 2.5° × 2.5° grid. The Niño-3.4 index is adapted from NOAA’s Climate Prediction Center (CPC). The SST dataset is from the Kaplan extended SST (version 2) at a 5° × 5° resolution (Kaplan et al. 1998).
Since this study focuses on the interannual time scale, a linear detrended processing is first applied to the above datasets prior to all the analyses. Then the fast Fourier transform (FFT) filter is used to remove the components beyond the interannual time scale (i.e., greater than 10 yr). This study stresses the interannual time scale in comparison with most of the previous studies in which unfiltered original data are used (e.g., Cong et al. 2007). The basic methods used include composite and correlation analyses. A Student’s t test is applied for checking the significance of the results. Because several various indices are adapted or defined in the study, Table 1 lists a brief introduction to them for the convenience of reference.
Description of various indices used in this study.
3. The correlations between the CEFs
Figure 1 (upper panels) compares the vertical distribution of the meridional wind component across the equator (averaged 2.5°S–2.5°N) over the Indian–western Pacific Ocean between the two reanalyses. Consistent in both reanalyses, there are five branches of cross-equatorial southerlies located near Somalia (37.5°–62.5°E, hereafter Somali CEF), the Bay of Bengal (82.5°–90°E), the South China Sea (102.5°–110°E), the Celebes Sea of the western Pacific (WP; 122.5°–130°E), and New Guinea (NG; 147.5°–152.5°E). The locations of the five branches exhibit a considerable consistence in both datasets, albeit with slightly stronger amplitudes in ERA-40. Previous studies have revealed that the latter three of the above five branches, north of Australia, are highly correlated with each other and may be an integral split by local topographies in the Maritime Continent. Correlation analyses are conducted for both reanalyses (see the appendix), and the results validate the substantial linkage among them. Thus, they are combined and referred to as the Australian CEF in the below analyses.
(a),(c) Vertical distribution of climatological summer (JJA) seasonal meridional wind component near the equator (m s−1) and (b),(d) horizontal distribution of surface winds over the Asian–Pacific monsoonal region. The climatology is calculated as the mean during 1958–2001. (left) Calculated based on the NCEP–NCAR reanalysis. (right) Calculated based on ERA-40. The horizontal solid lines in (a) and (c) stand for the vertical levels and the five rectangular boxes in (b) and (d) stand for the regions that are used to calculate the Somali, BOB, and three Australian CEFs, respectively.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
To obtain a direct vision at the CEFs, Fig. 1 (bottom panels) displays the spatial distribution of horizontal winds at the surface. Visually substantial cross-equatorial flows exist in the above-mentioned five regions (marked by five small boxes), and a substantial similarity is seen in both reanalyses. Previous studies suggested differences between the two reanalyses. Particularly, ERA-40 is more rational in describing the atmospheric mass flux between the hemispheres (Zhao and Li 2006). However no significant difference between the CEFs is seen here. One possible explanation is that both the datasets have been preprocessed by detrending and filtering, and the internal decadal gap in part responsible for their difference has been diminished.
Unlike the Somali CEF which has a maximum speed core around the 850-hPa level, both the BOB and the Australian CEF have a maximum at a relatively lower level near 925 hPa. Thus, the meridional wind component at 850 hPa is used to measure the strength of Somali CEF, while that at 925 hPa is used for the BOB and Australian CEF. Such a selection of vertical levels is slightly different from most of the previous studies (e.g., Cong et al. 2007; Zhu 2012), in which all the CEFs are defined on the same level, 925 hPa. An index is defined as the mean of the meridional wind components averaged over their individual regions shown in Table 2 for each CEF. The Australian CEF index is defined as the mean over the three regions north of Australia (Figs. 1b,d).
The geographic regions used to measure the strength of individual CEFs. SCS, WP, and NG indicate the three branches north of Australia, that is, the South China Sea CEF, the Celebes Sea CEF, and the New Guinea CEF, respectively. The mean of them are used to calculate the Australian CEF index.
Figure 2 displays the evolution of the CEF index anomalies in the past decades together with the climatological mean and standard deviation of the three CEFs (Somali, BOB, and Australian) in the two datasets. The strength of the CEFs exhibits an increase from the west to the east in terms of their locations. The standard deviation of the CEF index for the Somali, BOB, and Australian CEFs in the NCEP–NCAR reanalysis is 0.52, 0.52, and 0.91 individually, which exhibits an overall increase eastward. A similar feature is also seen in ERA-40 with the values of 0.36, 0.55, and 0.81, respectively. Visually, there exists substantial decadal variability in the CEFs. For example, for the Somali CEF there are high values around the late 1950s and early 1980s but lower values during the 1970s and the late 1990s.
A comparison of year-to-year evolution of CEF index anomaly. The panels from top to bottom are the Somali, BOB, and Australian CEFs, individually. (left) Calculated from the NCEP–NCAR reanalysis, and (right) calculated from ERA-40. The dashed lines are the values after one 11-yr running mean. The numbers on the upper-left corner in every panel indicate the std dev and the climatological average of the CEF index (m s−1), respectively.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
To check the interannual linkage between the CEFs, we filtered their individual decadal component and calculated their correlation coefficients by using the data in the period of 1970–2011. Only the data after year 1970 are used, because they are more creditable due to the application of satellite observations since then (Kistler et al. 2001). From Table 3, a negative correlation of the Australian CEF with the Somali CEF is significant in both datasets, suggesting a seesaw linkage between them. Besides, neither the Somali CEF nor the Australian CEF is significantly correlated with the BOB. In other words, the BOB CEF may vary independently relative to the Somali or the Australian CEF.
Correlation coefficient between the Somali, BOB, and the Australian CEFs calculated for the period 1970–2011. The bottom-left part is calculated from the NCEP–NCAR reanalysis, while the top-right part is from ERA-40. The boldface font indicates significance at the 95% level (with the correlation coefficient greater than 0.3 or less than −0.3 for the NCEP–NCAR reanalysis, greater than 0.35 or less than −0.35 for ERA-40). The values in brackets represent the results when the ENSO-related signals are removed.
To shed light on the variation of the above connection between the CEFs, Fig. 3 displays year-to-year evolution of the interannual correlation calculated in an 11-yr running window in both datasets. First the correlations are qualitatively consistent in two reanalyses. One significant negative correlation between the Somali and Australian CEFs exists in both reanalyses, albeit it is slightly stronger in ERA-40. The correlation between the Somali and the BOB CEFs is insignificant in both reanalyses. One inconsistence between the two datasets is a positive correlation prior to the mid-1960s in the NCEP–NCAR reanalysis, but it may be unrealistic due to the poor data quality during that time. Besides, decadal variation in the correlations is evident, although a discussion about it is beyond the scope of this study.
The interannual correlation between the Somali, BOB, the Australian CEFs in an 11-yr running window. The dashed lines represent statistical significance at the 95% confidence level.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
In view of the substantial negative correlation between the Somali and Australian CEFs, a seesaw index is defined by using the standardized mean strength of the Somali CEF minus that of the Australian CEF with the same weight for the two CEFs (Table 1). A higher (lower) seesaw index indicates the intensified (weakened) Somali CEF but weakened (intensified) Australian CEF. Figure 4a displays the seasonal cycle of the seesaw index. One may see that it peaks in summer [June–August (JJA)] and becomes opposite in winter along with seasonal annual cycle.
(a) The seasonal cycle of the seesaw index and summer (JJA) seesaw’s lead–lag correlation with the (b) Niño-3.4 index and (c) Indian Ocean DMI. The horizontal dashed lines represent the correlation coefficient significant at the 95% confidence level, that in (a) is for the correlation between the Somali and Australian CEFs.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
How the seesaw is connected to EASM and the lower-boundary forcing are investigated in the next sections. Because the seesaw peaks in boreal summer, all the below analyses are for the summer season (JJA) except for when specially clarified. Just the NCEP–NCAR reanalysis is used hereinafter in view of its longer record, although ERA-40 is more rational in describing atmospheric mass flux between the hemispheres (Zhao and Li 2006).
4. Connection to the East Asian summer climate
a. Rainfall and convection
Figure 5a shows the correlation of the Somali–Australian CEF seesaw index with rainfall in mainland China, which is derived from the gauged rainfall dataset at 160 stations within the country. Significantly positive correlations appear in the upper reach of the Yangtze River, Yellow River valley, and north China, while negative correlations are in the southern coast. The maximum positive correlation coefficient is 0.56 and the mean is 0.34, which both are significant at the 95% level. Furthermore, a statistics suggests that 51 (29) stations have the correlation significant above the 90% (95%) level, and a total of 10 stations have a positive coefficient over 0.4. This indicates an intensified EASM corresponding to a higher seesaw index.
Correlation of mainland China summer precipitation derived from the 160-station-gauged precipitation dataset with (a) the Somali–Australian CEF seesaw index, (b) simultaneous Niño-3.4 index, (c) Somali–Australian CEF seesaw index after the ENSO signals are removed, and (d) Niño-3.4 index in the previous winter [D(−1)JF]. The bigger and smaller black (gray) dots indicate a positive (negative) correlation significant at the 95% and 90% level, respectively. (e) Displays the distribution of 160 stations in China.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
To further check the robustness of connection between the seesaw and rainfall, a composite analysis is conducted. Based on the criterion that the seesaw index is greater (or smaller) than one positive (sign reversed) standard deviation, the 5 yr with the highest seesaw index (1973, 1984, 1989, 1998, and 2010) and the 5 yr with the lowest seesaw index (1972, 1979, 1982, 1987, and 1997) are identified for the period since 1970. Figure 6 (left panels) displays a comparison of the composite rainfall derived from the GPCP dataset. During the high seesaw index years, there is more rainfall in north China, but a deficiency in the southern coast, in agreement with the correlation above based on the station data (Fig. 5a). We also used the CPC Merged Analysis of Precipitation (CMAP) rainfall dataset (Xie and Arkin 1997) and obtained a similar result except for a slightly weakened significance in China (not shown). The consistency revealed in these different datasets suggests that the connection between the seesaw and East China rainfall is robust. Besides, above-normal (below normal) rainfall is located over the Maritime Continent, BOB, and the Indian Peninsula, along with decreased rainfall over the central to eastern tropical Pacific (Fig. 6c). Thus, there exist broad rainfall anomalies over the tropical Indian–Pacific Ocean and the Asian monsoon area linked to the seesaw.
Composite of rainfall (mm day−1) derived from the (left) GPCP dataset in the (a) high and (b) low Somali–Australian CEF seesaw index years and (c) their difference. (d)–(f) As in (a)–(c), but for the OLR (W m−2). In (c) and (f), positive, negative, and zero contours are drawn with solid, dashed, and thick lines, and the intervals of 1 and 10 and shading represent the significance over the 90% and 95% levels, respectively.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
The above rainfall anomalies are also reflected in convection seen in the OLR (right panels, Fig. 6). Climatologically there exist two convection activity centers (with smaller OLR), one locating over BOB and the other over the Philippines Seas. During the high seesaw index years, the convections over BOB and the Maritime Continent are stronger along with weakening over the equatorial central-eastern Pacific relative to the low seesaw index years (cf. Figs. 6d and 6e). This is particularly clear in their difference (Fig. 6f).
The above opposite connection in convections over the Maritime Continent and the equatorial central-eastern Pacific is reminiscent of the two major convection seesaws over the Indo-Pacific region, the Maritime Continent–Pacific convection oscillation (MPCO) and the Indo-Pacific convection oscillation (IPCO), found recently by Li et al. (2013). How the present CEF seesaw is connected with the MPCO and the IPCO is intriguing but unclear. Following Li et al. (2013), the MPCO and IPCO indices (MPCOI and IPCOI) are defined by the time series of the first and second leading EOF modes, respectively (Table 1). Figure 7 displays a comparison of the present CEF seesaw index with the MPCO and the IPCO indices. The CEF seesaw shows a high correlation with the MPCOI with a coefficient of 0.76, but no significant correlation with the IPCOI (with a coefficient of 0.16). This suggests that the CEF seesaw in part may be driven by the MPCO, but the underlying mechanism is unclear.
A comparison of the time series of the seesaw index, MPCOI, and IPCOI from 1970 to 2011. C. C. stands for the correlation coefficient with the seesaw index.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
Significant convection anomalies also exist over the southern subtropics and the equatorial Atlantic and Amazon. This suggests that the convection anomalies linked to the seesaw are not local or regional but global. The seesaw-related EASM and tropical convection connection is in agreement with Wang et al. (2001), which illustrated that the two major convection centers exhibit distinct difference when the ISM and the WNPSM is stronger or weaker.
b. Atmospheric circulation
In this subsection, the associated atmospheric circulation will be analyzed. Figure 8 shows the composite sea level pressure (SLP) and 850-hPa horizontal winds. For SLP, in both the high and low seesaw index years there are two regions with greater values in the southern subtropics, with centers locating over Australia and the Mascarene Islands individually. They reflect the two subsystems of Asian summer monsoon; the Australian high and the Mascarene high (Tao et al. 1962). In the high seesaw index years, the Australian high is weakened, meanwhile the Mascarene high is intensified, and vice versa. This is clearer from the composite difference (Fig. 8c). There are positive values in the southwest to the Mascarene Islands but negative values over Australia. This suggests that the seesaw reflects the opposite variation of the two highs, one intensifying and shifting to the west and the other weakening to some extent. That the Somali and Australian CEFs are connected to the two highs has been noted previously (Wang and Li 1982; Xue et al. 2003). Besides, the most significant negative values are over the tropical Indian Ocean and Maritime Continent as well as tropical Africa. Substantial negative anomalies are also seen over the Arab and India Peninsulas, which are consistent with the intensified ISM. In addition, there are positive anomalies over the subtropical western Pacific and eastern China. They reflect the enhancement of the WPSH, in agreement with the enhanced EASM.
As in Fig. 6, but for the (left) sea level pressure (hPa) and (right) wind vector at 850 hPa (m s−1) derived from the NCEP–NCAR reanalysis. (a),(b) Shading represents values greater than 1020 hPa, which are used to indicate the Mascarene and Australian subtropical highs. (c),(f) Shading represents significance over 95% level. Contour interval is 0.5 hPa in (c).
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
The above features are further seen in 850-hPa winds (Fig. 8f). There are anticyclonic and cyclonic anomalies over the locations of the Mascarene and Australian highs, respectively, along with the intensified southwesterly near the Somali coast and anomalous northwesterly on the east coast of Australia. Besides, there are intensified easterly trade winds in the tropical western Pacific and enhanced westerly in the Arabian Sea. These are correspondent to the intensified ISM and EASM.
Many previous studies investigated the correlation of the individual CEFs with the WPSH (e.g., Xue et al. 2003). Cong et al. (2007) suggested that the strength of the Somali CEF is well connected with the east–westward movement of the WPSH, while the SCS CEF is better related to the south–north fluctuation of the WPSH. Here we analyze the connection of the seesaw to the WPSH. Figure 9 (left panels) illustrates that during the high seesaw index years the WPSH is intensified with a westward extension (Fig. 9a). Besides another subtropical high, the Iran high, is also intensified. Opposite changes can be overall seen during the low seesaw index years (Fig. 9b). In addition to the subtropical anomalies, significant negative anomalies emerge over the tropical Indian Ocean.
As in Fig. 6, but for the geopotential heights at (left) 500 and (right) 100 hPa (gpdm). (a),(b) The shading indicates the climatological contours at 586 gpdm that are used to describe the climatological-mean location of the subtropical high over the western Pacific. (d),(e) Shading indicates the contours at 1680 gpdm used to describe the South Asian high, and (c),(f) the shading represents significance over the 95% level (gpdm). Contour interval is 0.5 and 1.0 in (c) and (f), respectively.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
The South Asia high (SAH) in the upper troposphere (~100 hPa) is another important member of the EASM system. The SAH is strengthened (weakened) and expands farther (shrinks) longitudinally during the high (low) seesaw index years (Figs. 9d,e,f). These changes are consistent with the enhancement (weakening) of the ISM or EASM. There is a similar height increase over the southern subtropics along with a decrease over the wide tropical region.
The connection of the seesaw with monsoonal circulation systems is additionally seen from its correlation with two EASM indices proposed by Li and Zeng (2002) and Wang et al. (2008), respectively. Table 1 (the last row) lists the detailed description of the indices. Albeit an intraseasonal difference, significant correlation is seen in July with the coefficient of −0.58 and 0.48, respectively, for the period from year 1970 to 2011. This opposite sign in correlation coefficient for the two EASM indices reflects a consistency considering that they are intrinsically anticorrelated (Wang et al. 2008).
The above results suggest that the enhancement of the EASM and ISM linked to the Somali–Australian CEF seesaw is exhibited consistently in many aspects, not only in rainfall but also in monsoonal circulation systems or monsoonal indices. Thus, the linkage is robust. Previous studies (e.g., Simmonds et al. 1999; Liu et al. 2009) suggest that both the Somali and Australian CEFs could affect the East Asian monsoon. We calculated the individual correlation of summer rainfall in mainland China with the Somali and Australian CEF as well as the seesaw index, respectively. The results (not shown) suggest a somewhat opposite connection to rainfall in northern China between the CEFs, in that a stronger Somali CEF contributes to intensified rainfall, while a stronger Australian CEF is associated with weakened rainfall. But their signals are not opposite in coastal southeastern China. This illustrates that their effects are not elusive and distinguishable from each other. Further, we calculated the station number and found more stations have significant correlation with the seesaw index than either CEF. Thus, it may bear more meaning to incorporate the CEF seesaw as a predictor rather than the individual CEFs.
5. Connection with SST
a. Connection with ENSO
Previous studies suggested various connection of CEFs to ENSO (Wang et al. 2001; Zhu and Chen 2002; Chen et al. 2005). This connection is also reflected in ENSO’s influence on the Mascarene and Australian highs. One preliminary result is that the Somali CEF tends to be weakened (strengthened), while the Australian CEF intensifies (weakens) during a warm (cold) ENSO episode. Zhu (2012) also illustrated that both the Somali and Australian CEFs are oppositely correlated with SST in the central-eastern tropical Pacific. Besides ENSO, the CEFs’ connections to SST in other oceans were also studied. Zhu (2012) found a negative correlation between the Somali CEF and the SST in the western tropical Indian Ocean and Arabian Sea. How the seesaw is connected to SST, that is, whether the connection is consistent with that derived from the individual CEFs, is intriguing.
First we compare composite SST anomaly (SSTA) prior to and following the summer seesaw and its evolution (Fig. 10). In the high seesaw index years, La Niña–like SSTA emerge simultaneously in the tropical central-eastern Pacific along with positive SSTA in the western Pacific and negative SSTA in the Indian Ocean. From the preceding winter to the following spring, SSTA features an evolution from mature El Niño to mature La Niña and then decaying La Niña. In contrast, in the low seesaw index years an overall opposite SSTA evolution is seen. This suggests that the seesaw is closely linked to ENSO.
(a)–(f) Composite SSTA from the preceding winter to the following spring for the high seesaw index years. (g)–(i) As in (a)–(f), but for the low seesaw index years. Contour interval is 0.25°C. The solid contour indicates 0. Shading indicates significance at the 95% level.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
To further examine the seesaw–ENSO relationship, we calculated the seesaw’s lead–lag correlation with the Niño-3.4 index. The 40-yr data series from 1970 to 2011 is used, although it may not be long enough to sample the intrinsic variability of ENSO (Wittenberg 2009). From Fig. 4b, a significant negative correlation (nearly −0.76) is seen persisting into the following winter [December–February (DJF)(+1)]. This suggests that a strong (weak) summer seesaw coincides with the developing phase of cold (warm) ENSO episodes as revealed in Fig. 10. However, it does not imply definitely that the seesaw leads ENSO, because the magnitude of the correlation is smaller than the ENSO’s lagged autocorrelation, which has a value of 0.87 from the developing summer (JJA) to the peak winter [DJF(+1)].
During the high (low) seesaw index summer, the weakened (strengthened) Australian CEF (Fig. 8) tends to decrease (increase) evaporation over the tropical eastern Indian Ocean and the Maritime Continent (Fig. 11a), accounting for the warming (cooling) of SST therein (Figs. 10c,i). This results in strengthening (weakening) of the Walker cell (Fig. 11b) along with easterly (westerly) anomalies near the surface. Then, cooler (warmer) SSTA in the tropical central-eastern Pacific can persist into the following winter (Figs. 10d,e,j,k).
(a) As in Fig. 6c, but for evaporation flux derived from ERA-40 (contour interval is 0.1 mm day−1), and (b) the Walker cell, anomaly indicated with the mean vector streamline (υ, ω) along the 2.5°S–2.5°N averaged longitudinal–vertical section. Here υ is the zonal component of horizontal wind (m s−1), while ω is pressure vertical velocity and scaled by 1000 (Pa s−1).
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
The seesaw–ENSO connection is consistent with the ENSO–East Asian summer rainfall. Summer rainfall in the Yangtze River valley and northeast China is negatively correlated with the simultaneous summer Niño-3.4 SSTA, but positively correlated with Niño-3.4 SSTA in the preceding winter (Figs. 5b,d). Thus, there is one correspondence among the summer seesaw, East Asian summer rainfall, the preceding winter El Niño, and the simultaneous La Niña.
Now that the seesaw is closely connected to ENSO, one intriguing question is whether the seesaw is just one phenomenon forced by ENSO. In other words, whether there exist seesaw signals independent of ENSO. To shed light on this, we had the ENSO-related signals removed and recalculated the correlation between the CEFs. The procedure to remove the ENSO-related signals is as in Li et al. (2006). First, the ENSO-induced anomaly is determined by regressing the time series of the interannual meridional wind component at each grid against the Niño-3.4 index time series. As a result, the ENSO signal for a particular month is a product of the ENSO-associated anomaly pattern and the ENSO index for that month. Last, the original interannual wind component minus the ENSO signal yields the ENSO-removed wind component. It should be noted that such a process just removes the linear signals. The values in brackets in Table 3 illustrate that the Somali–Australian CEF seesaw correlation becomes weaker but still pronounced. Thus, the seesaw may exist independent of ENSO.
Figure 12 displays the year-to-year evolution of the interannual correlation between the CEFs calculated in an 11-yr running window after the ENSO-related signals are removed. Obviously, similar features to Fig. 3 are seen except for slightly smaller values. A pronounced shift around 1965 is still there. When the ENSO-related signals are removed (in both the precipitation dataset and seesaw index), the correlation between the seesaw and the summer rainfall derived from the Chinese 160-station precipitation dataset (Fig. 5c) exhibits a substantial resemblance to that in the original datasets (Fig. 5a). This indicates that the connection of the seesaw with the EASM may exist independently, albeit ENSO modulates the strength of the correlation.
As in Fig. 3, but with the ENSO signals removed when calculating the CEF index.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
The independence of the seesaw–EASM is also reflected in the monsoonal circulation systems. After ENSO-related signals are removed, the composite SLP (Fig. 13a) still displays negative anomalies over the eastern Indian Ocean and north of Australia, corresponding to the weakened Australian high. The positive anomaly in the subtropical southwestern Indian Ocean reflects the westward shift and intensification of the Mascarene high, albeit being less significant. Also the opposite covariability between the Mascarene high and Australian high is seen linked to the seesaw. Similarly, at the 850-hPa wind the enhanced Somali CEF and weakened trade winds in the eastern Indian Ocean and southern tropical western Pacific north of Australia, along with the intensified easterly in the northern subtropical western Pacific, are still prominent, although the significant domain is much smaller. This means that the intensified WPSH and ISM are still correspondent with the seesaw, even without ENSO. Thus, the interannual seesaw between the Somali and Australian CEFs during boreal summer may be atmospherically and intrinsically variable within the Asian summer monsoon system, albeit the ENSO events tend to intensify it.
As in Figs. 8c and 8f, but with ENSO-related signals removed.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
b. Connection with IOD
The SSTA associated with the CEF seesaw (Figs. 10c–e,i–k) exhibits a dipolar structure in the tropical Indian Ocean, reminiscent of the well-known tropical IOD (Saji et al. 1999; Ashok et al. 2004). The IOD has been known to have a substantial influence on EASM (e.g., Yuan et al. 2008; Ding et al. 2010). Also, the lower-level wind anomaly linked to the seesaw (Figs. 8d–f) is very similar to the effect of the IOD as shown by Ashok et al. (2004, their Fig. 8) and Ummenhofer et al. (2011, their Fig. 3). The IOD and ENSO are synchronously correlated (e.g., Behera et al. 2006), but often considered to be independent modes of variability, although this is still a matter of debates. How the seesaw is connected to the IOD is thus an important issue.
We calculated the lead–lag correlation of the CEF seesaw with the dipole mode index (DMI; Table 1) reflecting the IOD and found the maximum negative correlation (−0.74) when the summer seesaw leads the IOD by one or two seasons (Fig. 4c). The magnitude of this correlation is close to that of the DMI’s lead–lag auto correlation from summer to fall (0.75). Thus, a negative IOD tends to coincide with the summer CEF seesaw. This may be physically reasonable. During summer with a higher seesaw index, the intensified southwesterly along the Somali coast causes an increase of evaporation (Fig. 11a) along with the intensified upwelling of cool water and the deepened thermocline (e.g., Izumo et al. 2008), which accounts for the cooling of SST in the western tropical Indian Ocean there. Meanwhile, weakened evaporation is seen in the eastern tropical Indian Ocean and offshore Sumatra (Fig. 11a) along with the intensified atmospheric ascent over the Maritime Continent (Fig. 11b). They together cause a negative-phase IOD in SST. Similar to its connection with ENSO, the correlation of the seesaw with the IOD is less significant when the IOD leads.
That the seesaw may exist independently is also seen when the IOD-related signals are removed. A similar method to the above to remove the ENSO signals is used to remove the IOD-associated variability. Calculations suggest that the negative correlation between the Somali and Australian CEFs remains unchanged (−0.41) after the IOD-associated signals are removed.
The above analyses suggest that the summer seesaw should not be explained just as one passive atmospheric response to the ENSO- or IOD-related SSTA. Instead, it may contribute to the formation of the IOD in the following fall. In view of its close connection with the EASM, the seesaw thus provides an additional benchmark for understanding the EASM.
6. Summary and discussions
In this study we first investigated the interannual correlation between the Somali, Bay of Bengal (BOB), and Australian cross-equatorial flows (CEFs), all of which are closely linked to the East Asian summer monsoon. No significant correlation was seen between the BOB CEF and the other two, but a significant negative correlation was illustrated between the Somali and Australian CEFs. The substantially negative correlation between the Somali and Australian CEFs suggests a seesaw oscillation. Thus, a seesaw index is defined with a greater (smaller) value indicating an intensified (weakened) Somali CEF but a weakened (intensified) Australian CEF. Then the connection of the seesaw with Asian summer monsoon is analyzed. The results suggest a substantial correlation, with a stronger seesaw corresponding to an intensified ISM and EASM and vice versa. Particularly, the seesaw essentially reflects the opposite fluctuation between the two permanent action centers in the Southern Hemisphere, the Mascarene high and the Australian high. During the higher (lower) seesaw index years, the Mascarene high moves westward (eastward), along with increased (decreased) convection in India. Meanwhile, the Australian high weakens (strengthens), along with the easterly (westerly) anomalies prevailing over the tropical western Pacific to the South China Sea. This favors enhancing the EASM. Subsequently the seesaw’s connection with ENSO and the Indian Ocean SST dipole (IOD) is explored. The results suggest that the higher (lower) seesaw tends to happen in the developing phase of La Niña (El Niño) and the negative IOD episode. When ENSO- or IOD-related signals are removed, both the seesaw itself and the seesaw–EASM connection are still significant. Thus, the CEF seesaw should not be explained as one passive response to ENSO or the IOD. Instead it contributes to the formation of the IOD.
Many previous studies have addressed the connection of various CEFs to the Asian summer monsoon, but primarily focus on individual CEFs. Here we illustrated that the CEFs are not all independent of each other and proposed a seesaw index to describe the opposite covariability between the Somali and Australian CEFs. This considers the effect of the combined CEFs rather than their individual one and thus bears more meaning. The present study suggests that the seesaw–EASM connection is closer than the ENSO–EASM connection, and the seesaw is also connected to the Maritime Continent–Pacific convection oscillation (MPCO) found recently by Li et al. (2013). Thus, the seesaw and its close connection to Asian summer monsoon may be intrinsic variability within the Asian Monsoon system. This contributes to understanding the interannual variability of summer rainfall in the Asian monsoonal region and provides additional insights into the EASM’s predictability.
It is well known that simulating and predicting the EASM is a big challenge even in current state-of-the-art atmospheric or climate system models. That the modeled monsoon does not reproduce the realistic one is one potential factor responsible for the biases. Now that the seesaw and seesaw–EASM connection are intrinsic variability, this provides one benchmark for understanding the models’ performance.
It should be pointed out that the present study only considers the interannual correlations between CEFs. Substantial decadal variations in this connection can also be seen (Figs. 3, 12). Thus, more comprehensive studies are needed to investigate their correlation in various time scales.
Finally, one may have noted a quasi-biennial oscillation signal in the CEF seesaw index (Fig. 7). One calculation of its power spectral confirms this signal (Fig. 14). How the CEF seesaw is connected to the tropospheric biennial oscillation (TBO) (e.g., Meehl and Arblaster 2002; Chang and Li 2000) is an intriguing issue and deserves an in-depth investigation, since the TBO was found to relate to the interaction between South Asian and Indo-Australian monsoons.
The power spectral distribution of the CEF seesaw index. The values above the dashed line are significant at the 95% level. The unit for frequency is yr−1, and the vertical coordinate is percentage of explained variance ratio.
Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00288.1
Acknowledgments
The authors thank three anonymous reviewers for constructive suggestions, which led to a significant improvement of the manuscript. This study is jointly supported by the strategic project of the Chinese Academy of Science (XDA11010401), the National Key Basic Research and Development (973) Program of China (2012CB417403), and the Special Research Fund for Public Welfare of China Meteorological Administration (GYHY201006022).
APPENDIX
Correlation Analyses between the CEFs North of Australia
There are three branches of cross-equatorial flows north of Australia, which are located in the South China Sea (102.5°–110°E), the Celebes Sea of the western Pacific (122.5°–130°E), and near New Guinea (147.5°–152.5°E), respectively (Fig. 1). One calculation of their correlations is conducted from both the NCEP–NCAR and the ERA-40 reanalyses (Table A1). All the correlations among them and their individual correlations with their combination, the Australian CEF, are over 0.6, significant at the 99% level. This indicates that they are not independent of each other, but an integral split by local topographies in the Maritime Continent.
Correlation coefficients between the three branches of CEFs north of Australia as well as their individual correlation with their combination, the Australian CEF. The calculation period is for 1970–2011. The values in the bottom-left corner are from the NCEP–NCAR reanalysis, while those in the top-right corner are from ERA-40. All the correlations are significant at the 99% level.
REFERENCES
Adler, R. F., and Coauthors, 2003: The version 2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 1147–1167, doi:10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.
Ashok, K., Z. Guan, N. H. Saji, and T. Yamagata, 2004: Individual and combined influences of ENSO and the Indian Ocean dipole on the Indian summer monsoon. J. Climate, 17, 3141–3155, doi:10.1175/1520-0442(2004)017<3141:IACIOE>2.0.CO;2.
Behera, S. K., J. J. Luo, S. Masson, S. A. Rao, H. Sakuma, and T. Yamagata, 2006: A CGCM study on the interaction between IOD and ENSO. J. Climate, 19, 1688–1705, doi:10.1175/JCLI3797.1.
Cadet, D. L., and G. Reverdin, 1981: Water vapor transport over the Indian Ocean during summer 1975. Tellus, 33, 476–487, doi:10.1111/j.2153-3490.1981.tb01773.x.
Cadet, D. L., and S. Greco, 1987: Water vapor transport over the Indian Ocean during the 1979 summer monsoon. Part I: Water vapor fluxes. Mon. Wea. Rev., 115, 653–663, doi:10.1175/1520-0493(1987)115<0653:WVTOTI>2.0.CO;2.
Chang, C.-P., and T. Li, 2000: A theory for the tropical tropospheric biennial oscillation. J. Atmos. Sci., 57, 2209–2224, doi:10.1175/1520-0469(2000)057<2209:ATFTTT>2.0.CO;2.
Chen, B., P. W. Guo, and Y. C. Xiang, 2005: Relationship between summer cross-equatorial flows and ENSO (in Chinese). J. Nanjing Inst. Meteor., 28, 36–43.
Cong, J., Z. Y. Guan, and L. J. Wang, 2007: Interannual (interdecadal) variabilities of two cross-equatorial flows in association with the Asian summer monsoon variations (in Chinese). J. Nanjing Inst. Meteor., 30, 779–785.
Ding, R., K. J. Ha, and J. Li, 2010: Interdecadal shift in the relationship between the East Asian summer monsoon and the tropical Indian Ocean. Climate Dyn., 34, 1059–1071, doi:10.1007/s00382-009-0555-2.
Findlater, J., 1966: Cross-equatorial jet streams at low levels over Kenya. Meteor. Mag., 95, 353–364.
Gruber, A., and A. F. Krueger, 1984: The status of the NOAA outgoing longwave radiation data set. Bull. Amer. Meteor. Soc., 65, 958–962, doi:10.1175/1520-0477(1984)065<0958:TSOTNO>2.0.CO;2.
Halpern, D., and P. M. Woiceshyn, 2001: Somali jet in the Arabian Sea, El Niño, and India rainfall. J. Climate, 14, 434–441, doi:10.1175/1520-0442(2001)014<0434:SJITAS>2.0.CO;2.
Han, S. Y., 2002: The time-space variation characteristics of cross-equatorial air flow and the relation to China summer rainfall and to tropical cyclone frequency in the Northwest Pacific (in Chinese). M.A. thesis, Nanjing Institute of Meteorology, 31–34.
Izumo, T., C. B. Montégut, J. J. Luo, S. K. Behera, S. Masson, and T. Yamagata, 2008: The role of the western Arabian Sea upwelling in Indian monsoon rainfall variability. J. Climate, 21, 5603–5623, doi:10.1175/2008JCLI2158.1.
Joseph, P. V., and S. Sijikumar, 2004: Intraseasonal variability of the low-level jet stream of the Asian summer monsoon. J. Climate, 17, 1449–1458, doi:10.1175/1520-0442(2004)017<1449:IVOTLJ>2.0.CO;2.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Kaplan, A., M. Cane, Y. Kushnir, A. Clement, M. Blumenthal, and B. Rajagopalan, 1998: Analyses of global sea surface temperature 1856–1991. J. Geophys. Res., 103, 18 567–18 589, doi:10.1029/97JC01736.
Kistler, R., and Coauthors, 2001: The NCEP–NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82, 247–267, doi:10.1175/1520-0477(2001)082<0247:TNNYRM>2.3.CO;2.
Li, J. P., and Q. C. Zeng, 2002: A unified monsoon index. Geophys. Res. Lett., 29, 1274, doi:10.1029/2001GL013874.
Li, J. P., and D. X. Hu, 2011: Chapter one. Air-Sea Interaction over the Joint Area of Asia and Indian-Pacific Ocean and its Impact on the Short-Term Climate Variation in China (in Chinese), Vol. 1, China Meteorological Press, 1–30.
Li, S., M. P. Hoerling, S. Peng, and K. M. Weickmann, 2006: The annular response to tropical Pacific SST forcing. J. Climate, 19, 1802–1819, doi:10.1175/JCLI3668.1.
Li, Y. J., J. P. Li, and J. Feng, 2013: Boreal summer convection oscillation over the Indo-western Pacific and its relationship with the East Asian summer monsoon. Atmos. Sci. Lett., 14, 66–71, doi:10.1002/asl2.418.
Li, Z. Z., and G. P. Lou, 1987: A study on the passages of cross-equatorial current during the southern monsoon (in Chinese). Sci. Atmos. Sin., 11, 313–319.
Liu, X. W., Z. B. Sun, D. H. Ni, W. J. Li, J. Y. Jia, and G. R. Tan, 2009: Connection of 105°E and 125°E cross-equatorial flow with the Southern and Northern Hemispheric circulations (in Chinese). Chin. J. Atmos. Sci., 33, 443–458.
Ma, L., T. Zhang, O. W. Frauenfeld, B. Ye, D. Yang, and D. Qin, 2009: Evaluation of precipitation from the ERA-40, NCEP-1, and NCEP-2 reanalyses and CMAP-1, CMAP-2, and GPCP-2 with ground-based measurements in China. J. Geophys. Res., 114, D09105, doi:10.1029/2008JD011178.
Mao, T. S., Q. C. Xu, and N. Y. Xu, 1990: The interrelation between the cross equatorial airflow near 105°E and the summer monsoon over Asia (in Chinese). Mar. Forecasts, 7, 1–7.
Meehl, G. A., and J. M. Arblaster, 2002: The tropospheric biennial oscillation and Asian–Australian monsoon rainfall. J. Climate, 15, 722–744, doi:10.1175/1520-0442(2002)015<0722:TTBOAA>2.0.CO;2.
Peng, W. G., and S. C. Jiang, 2003: The climatic characteristics of global cross-equatorial flow and its interannual variation in East Hemisphere (in Chinese). J. Trop. Meteor., 19, 87–93.
Ramesh Kumar, M. R., S. S. C. Shenoi, and P. Schluessel, 1999: On the role of the cross equatorial flow on summer monsoon rainfall over India using NCEP/NCAR reanalysis data. Meteor. Atmos. Phys., 70, 201–213, doi:10.1007/s007030050034.
Saha, K. R., and S. N. Bavadekar, 1973: Water vapour budget and precipitation over the Arabian Sea during the northern summer. Quart. J. Roy. Meteor. Soc., 99, 273–278, doi:10.1002/qj.49709942006.
Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363.
Shi, N., D. P. Shi, and M. L. Yan, 2001: The effects of cross-equatorial current on South China Sea monsoon onset and drought/flood in East China (in Chinese). J. Trop. Meteor., 17, 405–414.
Shi, N., G. L. Feng, J. Q. Gu, and D. J. Gu, 2007: The climatological variation of global cross-equatorial flow for the period of 1948-2004 (in Chinese). J. Trop. Meteor., 23, 326–332.
Simmonds, I., D. Bi, and P. Hope, 1999: Atmospheric water vapor flux and its association with rainfall over China in summer. J. Climate, 12, 1353–1367, doi:10.1175/1520-0442(1999)012<1353:AWVFAI>2.0.CO;2.
Tao, S. Y., and L. Chen, 1987: A review of recent research on the East Asian summer monsoon in China. Monsoon Meteorology, C.-P. Chang and T. N. Krishnamurti, Eds., Oxford University Press, 60–92.
Tao, S. Y., S. Y. Xu, and Q. Y. Guo, 1962: The characteristics of the zonal and meridional circulation over tropical and subtropical regions in eastern Asia in summer. Acta Meteor. Sin.,32, 91–103.
Ummenhofer, C. C., A. S. Gupta, Y. Li, A. S. Taschetto, and M. H. England, 2011: Multi-decadal modulation of the El Niño–Indian monsoon relationship by Indian Ocean variability. Environ. Res. Lett., 6, 034006, doi:10.1088/1748-9326/6/3/034006.
Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 2961–3012, doi:10.1256/qj.04.176.
Wang, B., and L. Ho, 2002: Rainy season of the Asian–Pacific summer monsoon. J. Climate, 15, 386–398, doi:10.1175/1520-0442(2002)015<0386:RSOTAP>2.0.CO;2.
Wang, B., R. G. Wu, and K. M. Lau, 2001: Interannual variability of the Asian summer monsoon: Contrasts between the Indian and western North Pacific–East Asian monsoon. J. Climate, 14, 4073–4090, doi:10.1175/1520-0442(2001)014<4073:IVOTAS>2.0.CO;2.
Wang, B., Z. W. Wu, J. P. Li, J. Liu, C.-P. Chang, Y. H. Ding, and G. X. Wu, 2008: How to measure the strength of the East Asian summer monsoon. J. Climate, 21, 4449–4463, doi:10.1175/2008JCLI2183.1.
Wang, H. J., and F. Xue, 2003: Interannual variability of Somali jet and its influences on the inter-hemispheric water vapor transport and on the East Asian summer rainfall (in Chinese). Chin. J. Geophys., 46, 11–20, doi:10.1002/cjg2.311.
Wang, J. Z., and M. C. Li, 1982: Cross-equator flow from Australia and monsoon over China (in Chinese). Sci. Atmos. Sin., 6, 1–10.
Wang, W. P., and X. Q. Yang, 2008: Variation of Somali Jet and its impact on East Asian summer monsoon and associated China rainfall anomalies (in Chinese). Sci. Meteor. Sin., 28, 139–146.
Wittenberg, A. T., 2009: Are historical records sufficient to constrain ENSO simulations? Geophys. Res. Lett., 36, L12702, doi:10.1029/2009GL038710.
Xie, P., and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78, 2539–2558, doi:10.1175/1520-0477(1997)078<2539:GPAYMA>2.0.CO;2.
Xue, F., H. J. Wang, and J. H. He, 2003: Interannual variability of Mascarene high and Australian high and their influences on summer rainfall over East Asia. Chin. Sci. Bull., 48, 492–497, doi:10.1007/BF03183258.
Yuan, Y., H. Yang, W. Zhou, and C. Li, 2008: Influences of the Indian Ocean dipole on the Asian summer monsoon in the following year. Int. J. Climatol., 28, 1849–1859, doi:10.1002/joc.1678.
Zhang, R. H., 2001: Relations of water vapor transport from Indian monsoon with that over East Asia and the summer rainfall in China. Adv. Atmos. Sci., 18, 1005–1017.
Zhao, Y. F., and J. P. Li, 2006: Discrepancy of mass transport between the Northern and the Southern Hemisphere among the ERA-40, NCEP/NCAR, NCEP-DOE AMIP-2, and JRA-25 Reanalysis. Geophys. Res. Lett., 33, L20804, doi:10.1029/2006GL027287.
Zhu, Q. G., J. H. He, and P. X. Wang, 1986: A study of circulation differences between East-Asian and Indian summer monsoons with their interaction. Adv. Atmos. Sci., 3, 466–477, doi:10.1007/BF02657936.
Zhu, Y. F., and L. X. Chen, 2002: The relationship between the Asian/Australian monsoon and ENSO on a quasi-four-year scale. Adv. Atmos. Sci., 19, 727–740, doi:10.1007/s00376-002-0012-1.
Zhu, Y. L., 2012: Variations of the summer Somali and Australia cross-equatorial flows and the implications for the Asian summer monsoon. Adv. Atmos. Sci., 29, 509–518, doi:10.1007/s00376-011-1120-6.
