Abstract

In austral summer, El Niño–Southern Oscillation (ENSO) covaries with the Indian Ocean Basin Mode (IOBM) and with the southern annular mode (SAM). The present study addresses how the IOBM and the SAM modulate the impact of ENSO on Australia. The authors show that the modulating effect of the SAM is limited; in particular, the SAM does not modify the ENSO teleconnection pattern. However, the IOBM extends ENSO-induced convection anomalies westward over northern Australia and over the eastern Indian Ocean, whereby extending the ENSO tropical teleconnection to the northwest of Australia. The IOBM also generates an equivalent-barotropic Rossby wave train through convection anomalies over northern Australia. The wave train shares an anomaly center over the Tasman Sea latitudes with the Pacific–South American (PSA) pattern, shifting the anomaly center of the PSA pattern to within a closer proximity to Australia. There is a strong asymmetry in the IOBM modulating effect. During an IOBM negative phase, which tends to coincide with La Niña events, the rainfall increase is far greater than the reduction during a positive IOBM phase, which tends to coincide with El Niño events. This modulation asymmetry is consistent with an asymmetry in the ENSO–rainfall teleconnection over Australia, in which the La Niña–rainfall teleconnection is stronger than the El Niño–rainfall teleconnection. This asymmetric ENSO–rainfall teleconnection ensures a higher coherence of northern Australia convective anomalies with La Niña or with a negative phase of the IOBM, hence a greater modification of the PSA pattern, underpinning the asymmetric modulating role of the IOBM.

1. Introduction

El Niño–Southern Oscillation (ENSO) triggers tropical and extratropical teleconnections through its tropical sea surface temperature (SST) and associated convection variations. The tropical teleconnection is an equatorially trapped, deep baroclinic response to the diabatic convective heating anomalies induced by the tropical SST anomalies (Gill 1980). The atmospheric manifestation of this direct tropical teleconnection is the Southern Oscillation (SO), of which the impact is mainly confined to near-tropical latitudes. The tropical diabatic heating anomalies also excite equivalent barotropic Rossby wave trains that propagate into the extratropics (Hoskins and Karoly 1981). Such tropical and extratropical responses also operate in association with the Indian Ocean dipole (IOD; Saji and Yamagata 2003; Cai et al. 2009b, 2011) during austral winter and spring, contributing to drought in Indonesia and Australia (Ashok et al. 2003; Cai et al. 2005; Meyers et al. 2007), bushfires over southeast Australia (Cai et al. 2009a), and floods in eastern Africa (Black et al. 2003; Zubair et al. 2003; Behera et al. 2005). In particular, the equivalent barotropic Rossby waves emanating from one or both poles of the IOD are a major pathway whereby the IOD impact is conveyed to mid- to high-latitude regions such as China and Australia (Cai et al. 2011).

Focusing on the impact on Australia, during austral winter [June–August (JJA)], the impact of ENSO on Australia is confined to eastern Australia through the tropical teleconnection (Cai et al. 2011). The IOD, largely independent from the ENSO in this season, generates its extratropical teleconnection by exciting equivalent barotropic Rossby wave trains from the eastern pole, with a pressure anomaly center south of Australia influencing climate variability. In austral spring [September–November (SON)] when ENSO and the IOD are highly coherent, the impact of ENSO on southern Australia is conducted through its coherence with the IOD, as the anomaly centers of the Pacific–South American (PSA) and the Pacific–North American (PNA) patterns are located remote from Australia and do not have an impact pathway (Cai et al. 2011). They showed that equivalent barotropic Rossby wave trains emanate from the two poles of the IOD, referred to as the eastern Indian Ocean and western Indian Ocean wave trains. These wave trains share a pressure center south of Australia, slightly westward compared with that during JJA, and are the major avenue for the IOD to impact southern Australia. The impact of ENSO in this season is conducted through its coherence with the IOD (Cai et al. 2011). This explains why the pattern of ENSO impact in SON tends to be broad (McBride and Nicholls 1983; Nicholls 1985), more so than that in December–February (DJF). Further, the impact through the Indian Ocean pathway is more prominent during positive IOD (pIOD) events than during negative IOD (nIOD) events (i.e., the impact is asymmetric about the positive and negative phases; Cai et al. 2012a).

In austral summer (DJF), the dominant mode of SST variability in the Indian Ocean, the Indian Ocean Basin Mode (IOBM; Chambers et al. 1999; Klein et al. 1999), varies coherently with ENSO (Xie et al. 2009; Du et al. 2009), reminiscent of the relationship between ENSO and the IOD in the austral spring season (SON). These previous studies have revealed that the IOBM is a response to ENSO through the “atmosphere bridge” involving winds and turbulent fluxes (Klein et al. 1999). During El Niño events, the convection center shifts away from the western Pacific and the eastern Indian Ocean. This leads to decreasing cloud cover and increasing net heat flux into the ocean, conducive to warm SST anomalies. In addition, easterly wind anomalies develop over the eastern Indian Ocean, which weaken the climatological winds, further enhancing the warming. In turn, these wind anomalies induce off-equatorial downwelling oceanic Rossby waves, which propagate westward. These waves then generate warming in the western part of the basin through thermocline coupling in this season (Xie et al. 2002; Cai et al. 2005), leading to a basinwide warming that persists well after the mature phase of ENSO: the so-called Indian Ocean capacitor effect (Xie et al. 2009). The reverse is generally true for the negative phase of the IOBM.

How does the IOBM modulate the impact of ENSO during austral summer? Is there an asymmetry in the modulating effect? These issues are largely unexplored, although Taschetto et al. (2011) used a model to study the individual and combined impact of ENSO and the IOBM. They found that in addition to modifying the ENSO-induced tropical teleconnection, the positive phase the IOBM excites a pair of barotropic anomalies in the Indian Ocean extratropics, one of which is an anomalous anticyclone in the Great Australian Bight. They show that through the “capacitor effect” (Xie et al. 2009), the IOBM extends the impact of ENSO from January to March. However, they did not address whether the modulation of the IOBM positive and negative phases on ENSO is symmetric.

In this study, we discuss the impact of the IOBM and ENSO and their interaction (section 3). We show that there is no significant anomalous anticyclone circulation over the Great Australian Bight region associated with a positive IOBM, in contrast to the modeling result of Taschetto et al. (2011); instead, for a positive phase of the IOBM, there is a tendency for a cyclonic center over the Tasman Sea latitudes, which extends the ENSO–rainfall teleconnection southward along the east coast of Australia (section 4). Further, there is a strong asymmetry in the role of the IOBM during positive and negative phases (section 5). In the DJF season, ENSO also covaries with the southern annular mode (SAM; L’Heureux and Thompson 2006). How does such coherence modulate the impact of ENSO? To answer this question, we also explore whether ENSO impacts are modulated by the SAM (section 6). A summary is given in section 7.

2. Data and methods

We use the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalyses for analyzing mean sea level pressure (SLP), 200-hPa geopotential height (GPH), and outgoing longwave radiation (OLR; Kalnay et al. 1996). Based on dataset availability, we constrain our analysis to the period 1949–2011. We define the year of an austral summer (DJF) as the year of the January. To diagnose effects on Australian rainfall, a 0.05° gridded dataset using quality-controlled station data from the Australian Bureau of Meteorology is used (Jones et al. 2009). Gridded SST, obtained from the Hadley Centre Sea Ice and Sea Surface Temperature dataset analyses (HadISST; Rayner et al. 2003), are used to produce an oceanic ENSO index, Niño-3.4, which is the average SST anomalies over 5°S–5°N, 170°–120°W. The Indo-Pacific spatial pattern of Niño-3.4 regressed onto SST is displayed in Fig. 1a. An index for the IOBM is obtained using the same SST dataset, calculated using the first empirical orthogonal function (EOF) of detrended SST in the Indian Ocean domain of 40°S–30°N, 30°–140°E. The first EOF mode accounts for 27.4% of the total variance, and is well separated from the subsequent modes according to the North et al. (1982) criterion. The mode shows a warm phase (Fig. 1c), corresponding to the positive phase of the index. ENSO is highly correlated with IOBM (Fig. 2a), as reflected by the strong resemblance of their anomaly patterns (cf. Figs. 1a and 1c). Our SAM index is obtained from Dr. G. Marshall’s website (http://www.antarctica.ac.uk/met/gjma/sam.html), and was produced using observational station pressure data only (Marshall 2003). Because of the reliance on station data, this SAM index is limited to 1958 onward; therefore, analyses referring to the SAM are restricted to the period 1958–2011. It should be noted that ENSO has a correlation with the SAM that is statistically significant at the 95% confidence level; however, the IOBM and the SAM are uncorrelated.

Fig. 1.

One standard deviation anomaly pattern of SST anomalies associated with (a) Niño-3.4, (b) Niño-3.4 with the IOBM removed, (c) the IOBM, and (d) the IOBM with Niño-3.4 removed. Green contours show the statistically significant correlation coefficients at the 95% confidence interval (in this And all figures with green contours). IOBM|Niño-3.4

Fig. 1.

One standard deviation anomaly pattern of SST anomalies associated with (a) Niño-3.4, (b) Niño-3.4 with the IOBM removed, (c) the IOBM, and (d) the IOBM with Niño-3.4 removed. Green contours show the statistically significant correlation coefficients at the 95% confidence interval (in this And all figures with green contours). IOBM|Niño-3.4

Fig. 2.

(a) Diagram showing the correlation between IOBM, Niño-3.4, and the SAM indices. Correlations ≥ 95% significance are highlighted in red. Scatterplots of (b) IOBM; and (c) SAM vs Niño-3.4, with the linear regression coefficient (slopes), correlations, and p values of the linear fits conducted using all samples, only positive, and only negative values of Niño-3.4.

Fig. 2.

(a) Diagram showing the correlation between IOBM, Niño-3.4, and the SAM indices. Correlations ≥ 95% significance are highlighted in red. Scatterplots of (b) IOBM; and (c) SAM vs Niño-3.4, with the linear regression coefficient (slopes), correlations, and p values of the linear fits conducted using all samples, only positive, and only negative values of Niño-3.4.

Our main technique of analysis is linear regression and correlation for diagnosing the relevant signals associated with an index. We use partial regression and correlation techniques to isolate one index from another. Partial regression (correlation) involves computation of the linear regression (correlation) of a predictand upon a predictor, after the linear relationship with a second predictor has been removed from both the predictand and predictor. For example, we use the notation “IOBM|Niño-3.4” to indicate the effect of the IOBM on a specific field without the influence of Niño-3.4 (i.e., with the ENSO covariance removed). For the opposite case, where the influence of IOBM is removed from Niño-3.4, we would refer to this as “Niño-3.4|IOBM.” Figures 1b and 1d show the SST variability patterns after removing the IOBM signal from ENSO, and after removing the ENSO signal from the IOBM, respectively. It must be noted that removing the influence of one index from another removes all coherent portions, whether related or incidental. Because regression analysis provides information of the sensitivity of a predictand field to a predictor, but with no information on their amplitudes, to compare the relative effects we multiply the regression and partial regression coefficient fields by the predictor’s standard deviation. Furthermore, to highlight the significance of a relationship we contour the 95% statistically significant correlation coefficient.

To investigate possible asymmetry, regression coefficients are also calculated using samples containing only positive index values (32 yr) or only negative index values (31 yr) and compared with that calculated using all samples. The corresponding correlations are also calculated to indicate whether statistical significance has been achieved above the 95% confidence level. Figure 2b shows an example using the IOBM and Niño-3.4 time series. An asymmetry exists, as the relationship between the IOBM and positive Niño-3.4 (El Niño) values is stronger than that between the IOBM and negative Niño-3.4 (La Niña) values (Fig. 2b). Similarly, the relationship between the SAM and El Niño is stronger than that between the SAM and La Niña, although in this case the difference is very slight (Fig. 2c). When such analyses are conducted using climate fields and climate indices, patterns of regression coefficients are constructed. A comparison of the regression coefficient patterns is then made, in terms of per unit change of the positive and negative index values, so as not to be influenced by the amplitude skewness of each index. For a clearer illustration, in the color legends we always use blue to indicate increased rainfall, cold anomalies, and enhanced convection.

3. The tropical teleconnection of the IOBM and the ENSO

As discussed in the introduction, many studies have examined dynamical processes associated with the IOBM (Chambers et al. 1999; Klein et al. 1999; Xie et al. 2009; Du et al. 2009). Focusing on its impact, we point out that SST anomalies in the Indian Ocean, although of the same sign, are not spatially uniform: warm anomalies in the off-equatorial Indian Ocean are greater than in the equatorial region, particularly to the south of the equator, where a zone of a maximum warming extends northwestward toward the equator (Fig. 1c). In this season, a southern Indian Ocean dipole (SIOD; Venegas et al. 1997; Behera and Yamagata 2001; Reason 2001; Fauchereau et al. 2003) operates, largely forced by the atmosphere (Fauchereau et al. 2003). Consequently, SST off Sumatra–Java and off northwest Australia is less warm than that of the west Indian Ocean.

For discussion purposes, we show anomaly patterns associated with ENSO in Figs. 3a–c. The ENSO teleconnection features a SO pattern, which in the Pacific encompasses a baroclinic, tropically trapped response in the tropical eastern Pacific (the tropical teleconnection). During El Niño, cool SST suppresses convection in the western Pacific, and the South Pacific convergence zone (SPCZ) moves toward the equator (Fig. 3a) (e.g., Folland et al. 2002; Vincent et al. 2011; Cai et al. 2012b).

Fig. 3.

(left to right) One standard deviation anomaly pattern of OLR (W m−2), 200-hPa geopotential height (GPH, m), and Australian rainfall (mm) associated with (top to bottom) Niño-3.4, Niño-3.4 with the IOBM removed, the IOBM with Niño-3.4 removed, and the IOBM.

Fig. 3.

(left to right) One standard deviation anomaly pattern of OLR (W m−2), 200-hPa geopotential height (GPH, m), and Australian rainfall (mm) associated with (top to bottom) Niño-3.4, Niño-3.4 with the IOBM removed, the IOBM with Niño-3.4 removed, and the IOBM.

The correlation between the IOBM and ENSO is 0.8 (Fig. 2a), indicating that approximately 60% of the IOBM variance is coherent with ENSO. When coherent with ENSO, the weaker warm SST anomalies (or relative cooling compared with the anomalies to the west) are seen off the Sumatra–Java and northwest Australia coasts (Fig. 1c). Along the southern equatorial Indian Ocean (7°–17°S), the west (60°–80°E) minus east (80°–100°E) SST gradient associated with the pattern is 0.25°C, during the analysis period. This facilitates a westward extension of the suppressed convection over the western Pacific into the eastern Indian Ocean, modifying the ENSO tropical teleconnection, as the convection center shifts to the highest SST region. The associated convection signal over Australia (Fig. 3j), compared with that associated with Niño-3.4 (Fig. 3a), extends slightly south. Locally, over waters off the Sumatra–Java and northwest Australia coasts, a situation arises in which reduced convection is associated with a higher local SST. This highlights the fact that it is the SST gradient, not the absolute SST, that is important.

Once the IOBM coherent signal is removed (Figs. 3d–f), the ENSO–rainfall teleconnection is confined to eastern Australia. Consistently, OLR anomalies over the eastern Indian Ocean and northern Australia are eliminated. These features suggest that the IOBM facilitates a broader impact from ENSO. This facilitating mechanism is clearer when the coherent ENSO anomalies are removed (Figs. 3g–i). After removal of ENSO, the weaker warm SST anomalies off the Sumatra–Java and northwest Australia coasts are still present, as are OLR and rainfall anomalies over northwestern Australia (Figs. 3g,i). There is also an OLR anomaly extending from the southern Indian Ocean passing south of Australia. This OLR anomaly appears to be a result of subsidence associated with the downward branch of the Hadley and Ferrel cells.

4. Modulation of the IOBM on the extratropical teleconnection of ENSO

Extratropical teleconnection features produced by ENSO is observed in anomaly patterns of 200-hPa geopotential height, which are referred to as the PSA and PNA pattern, in accordance with the Hoskins and Karoly (1981) model (Fig. 3b). The PSA pattern has a height anomaly over the area from the central Pacific to New Zealand. Once the IOBM-related anomalies are removed, the height anomaly over the Tasman Sea area (or New Zealand) is no longer apparent (Fig. 3e).

In contrast to the Taschetto et al. (2011) modeling study, there is only a weak indication of a wave train pattern in the Indian Ocean extratropics associated with the IOBM. In their model, there is a prominent anomalous anticyclone pattern over the Great Australian Bight, in response to a warm phase of the IOBM. In our analysis there is a hint of such a response, with a positive geopotential height anomaly situated to the west of the Great Australia Bight, or over the ocean west of Perth, Australia (Fig. 3h), but the response is not statistically significant. As expected, with this insignificant height anomaly, there is little impact on southern Australia rainfall (Figs. 3i,l). Another marked difference to the Taschetto et al. (2011) result is the presence of a strong cyclonic anomaly southeast of Tasmania (Fig. 3h) in our study, which is absent in theirs.

We propose that the geopotential height anomaly over New Zealand near the Tasman Sea is a result of Rossby wave trains originating from convective anomalies over northern Australia. After ENSO is removed, there are still OLR anomalies (Fig. 3g) over the Australian sector: a positive residual OLR anomaly over northern Australia extending from the tropical northwest, and a band of negative OLR anomaly extending from extratropical southern Indian Ocean passing southern Australia. We suggest that the IOBM, by modifying the spatial extent of the tropical convective anomalies over northern Australia, contributes to the height anomaly in the Tasman Sea, and modulates the extratropical response to ENSO, as we discuss below.

The IOBM-induced wave train appears to be incorporated onto the ENSO-induced PSA pattern, modifying the structure of the PSA so that the pressure center is situated closer to Australia (cf. Figs. 3b,e). A large portion of PSA variance is not observed in the absence of the IOBM (Fig. 3e). Overall, the pattern is consistent with the idea that the PSA pattern should be interpreted as the net resultant wave train pattern in response to tropical convective anomalies, rather than an individual anomaly center (Cai et al. 2011).

Comparing Australian rainfall regressed onto Niño-3.4|IOBM and IOBM|Niño-3.4 (Figs. 3f and 3i, respectively), with their original counterparts (Figs. 3c and 3l, respectively), we see several features. First, in the absence of coherence with the IOBM, the ENSO–rainfall teleconnection only operates over northeastern Australia, with little signal along Australia’s east coast south of 20°S (Fig. 3f). Likewise, without coherence with ENSO, the IOBM displays virtually no influence along Australia’s east coast (Fig. 3i). In other words, only in the presence of coherence with the IOBM, does the ENSO–rainfall teleconnection extend to Australia’s southern east coast (cf. Figs. 3c and 3f). Second, only in the presence of an influence from ENSO, does the IOBM display an impact on Australia’s southern east coast. These features highlight the primary forcing of ENSO, and the secondary modulating role of the IOBM.

The modulating effect of the IOBM is conducted by the IOBM-induced wave train over the Tasman Sea, which shifts the low pressure center of the PSA westward. With a closer proximity of a PSA anomaly center to Australia, the associated offshore–onshore flows are conducive for a rainfall impact along the southern east coast. Without the modulating effect from the IOBM, the ENSO impact along Australia’s east coast is mainly confined to the northern portion (Fig. 3f).

To confirm that it is indeed the convection anomalies near northern Australia that drive the low pressure center near the Tasman Sea, Fig. 4 shows a regression of gridpoint OLR anomalies onto a time series of 200-hPa geopotential height, averaged over the Tasman Sea region (47°–32°S, 150°–175°E). The results support this notion that the height anomaly over the Tasman Sea is associated with an enhanced convection (negative OLR anomalies) over the tropical western Pacific–northern Australia region (Fig. 4a), and such an association exists even after the removal of ENSO (Fig. 4b). This linkage to tropical convective anomalies is consistent with findings of previous diagnostic and modeling studies, which showed that diabatic heating in the vicinity of the Maritime Continent can excite a Rossby wave train that propagates into the Southern Hemisphere (Jin and Hoskins 1995; DeWeaver and Nigam 2004).

Fig. 4.

One standard deviation anomalies of gridpoint OLR associated with 200-hPa geopotential height over the Tasman Sea region (47°–32°S, 150°–175°E): (a) raw data and (b) after removal of variance coherent with ENSO.

Fig. 4.

One standard deviation anomalies of gridpoint OLR associated with 200-hPa geopotential height over the Tasman Sea region (47°–32°S, 150°–175°E): (a) raw data and (b) after removal of variance coherent with ENSO.

The band of reduced convection (positive OLR anomalies) extending from the extratropical southern Indian Ocean passing southern Australia cannot be generated by SST anomalies associated with the SIOD, as the positive OLR anomalies overlay a band of positive SST anomalies. Instead, it is most likely a consequence of increased subsidence associated with the enhanced tropical convection over the western Pacific–northern Australia regions. This is in turn consistent with the notion that the SIOD is induced by atmospheric processes.

5. Asymmetry in the modulation of IOBM on ENSO teleconnections

In sections 3 and 4, we show that in austral summer, the involvement of the Indian Ocean SST anomalies significantly modulates the ENSO–rainfall teleconnection on Australia. The eastern Indian Ocean may be regarded as part of the region with suppressed convection during El Niño, thus extending the impact of ENSO into the northwestern region of Australia. The associated convection anomalies over the tropical eastern Indian Ocean–North Australia also induce Rossby wave trains, which share an anomaly center with the ENSO-induced PSA pattern in the southwestern Pacific. However, the resultant center is situated closer to Australia when compared with that without an IOBM-coherent anomaly, enhancing the impact of ENSO on the southern Australia east coast.

Thus, similar to the situation in SON, in which much of the ENSO impact on southern Australia is conducted through the Indian Ocean (Cai et al. 2010), the impact of ENSO on northwest Australia and the east coast of southern Australia is modified by the IOBM. So far our analysis assumes a symmetric impact of ENSO and the IOBM with respect to their positive and negative phases. Recent studies have shown that a strong asymmetry exists, in terms of the impact from ENSO (Cai et al. 2012a). We therefore examine if the modulation by the IOBM is asymmetric.

a. Asymmetry in ENSO–Australian rainfall teleconnections

Previous studies have shown that much of the northern Australian rainfall–ENSO teleconnection is achieved by a rainfall increase during La Niña (Cai et al. 2010, 2012a; Cai and van Rensch 2012). Figure 5 plots regression (and correlation) coefficients of gridpoint rainfall anomalies onto Niño-3.4 using samples with a positive (Fig. 5a) and a negative (Fig. 5b) Niño-3.4 value, separately. These coefficients are similarly obtained as in Figs. 2b or 2c but conducted for each grid point. The results show that La Niña–induced northeastern Australia rainfall increase is proportionate to the amplitude of La Niña, whereas El Niño–induced rainfall reduction is generally not dependent on the amplitude of El Niño. This highlights the systematic difference between the two phases; in particular, the small regression coefficients obtained using the positive Niño-3.4 index.

Fig. 5.

One standard deviation anomalies of Australia rainfall associated with (a) positive Niño-3.4 values (El Niño) and (b) negative Niño-3.4 values (La Niña). Contour colors have been adjusted, so blue (red) represents rainfall increase (decrease).

Fig. 5.

One standard deviation anomalies of Australia rainfall associated with (a) positive Niño-3.4 values (El Niño) and (b) negative Niño-3.4 values (La Niña). Contour colors have been adjusted, so blue (red) represents rainfall increase (decrease).

The process underpinning this asymmetry is a zonal movement along the equator, and a meridional shift of the tropical convection center from the climatological mean position (Cai et al. 2010). During El Niño, the convection center (i.e., the SPCZ) moves eastward and to the equator, away from Australia. Once away from Australia, a stronger El Niño or a farther eastward movement would not further intensify the rainfall reduction. During La Niña, the convection center moves westward and expands away from the equator to encompass northern Australia. Once in place over northern Australia, the impact is proportionate to the La Niña amplitude. Thus, during El Niño, convection anomalies of the SPCZ are completely remote from Australia, whereas SPCZ variability during La Niña operates over northern Australia.

b. Asymmetry in the modulation by the IOBM

The modulation of ENSO impact by the IOBM is also asymmetric. To illustrate this, we regress gridpoint anomalies onto the IOBM using samples with a positive and negative phase, separately (Fig. 6). We start with an examination in the presence of ENSO-related anomalies. For the positive IOBM phase (left column, Fig. 6), the Indian Ocean SST anomalies are fairly uniform (Fig. 6a). Representative of the strong positive IOBM–El Niño coherence (Fig. 2b), convection anomalies are small over northern Australia (Fig. 6c), and there is only a weak impact on northern Australia rainfall (Fig. 6g). However, strong OLR anomalies arising from movement of the intertropical convergence zone (ITCZ) are present in the Northern Hemisphere, indicating enhanced convection along the central and eastern equator and suppressed convection to the north. The associated PNA anomalies are larger than the PSA signals. There are only weak height anomalies in the vicinity of the Tasman Sea (Fig. 6e).

Fig. 6.

(top to bottom) One standard deviation anomalies of SST (°C), OLR (W m−2), 200-hPa geopotential height (GPH, m), and Australian rainfall (mm), associated with (left) positive IOBM values and (right) negative IOBM values. Contour colors have been adjusted for simple comparison (see text).

Fig. 6.

(top to bottom) One standard deviation anomalies of SST (°C), OLR (W m−2), 200-hPa geopotential height (GPH, m), and Australian rainfall (mm), associated with (left) positive IOBM values and (right) negative IOBM values. Contour colors have been adjusted for simple comparison (see text).

These features are in sharp contrast to those associated with the negative phase, in which a stronger coherence with the southern tropical Pacific SST is seen (Fig. 6b). Strong convection anomalies are seen over northern Australia, with a tropical teleconnection with large rainfall anomalies extending into Australia from the northwest (Fig. 6d). Despite being less coherent with the Pacific La Niña, the negative IOBM-induced wave train with a pressure center situated over the Tasman Sea latitudes is rather prominent. The pressure center extends eastward, deep into the Pacific, reminiscent of the PSA pattern (Fig. 6f). This is likely associated with subsidence due to the convergence of the Hadley and Ferrel cells and is consistent with the decreased convection (positive OLR anomaly) south of Australia. The onshore flows associated with the high pressure center (corresponding to the negative IOBM) are conducive to a rainfall increase over eastern Australia (Fig. 6h).

Thus, much of the modulation by the IOBM on the impact of ENSO, seen in Fig. 3, is conducted during the negative phase of the IOBM. Since Fig. 6 contains ENSO-induced anomalies that are coherent with the IOBM, we construct anomaly patterns that are linearly independent from ENSO. To remove ENSO signals, we isolate the positive and negative IOBM years for each gridpoint and Niño-3.4 value and then use partial regression to ensure a more complete removal (rather than using an all-sampled linear regression). This is desirable because the relationship between anomalies and each of the two IOBM phases is not linear: El Niño is more coherent with the positive IOBM than La Niña with the negative IOBM (Fig. 2b).

Anomaly patterns associated with the IOBM incoherent with ENSO (Fig. 7) generally resemble those shown in Fig. 6. For the negative phases, there is a clear SIOD pattern in the southern Indian Ocean associated with the negative IOBM (Fig. 7b). This is not as clear during the positive phases, suggesting that the relationship between the SIOD and the IOBM or between the SIOD and ENSO is not symmetric. The dynamics behind these asymmetric relationships, however, is beyond the scope of the present study.

Fig. 7.

As in Fig. 6, but with Niño-3.4 linearly removed for the positive and negative phases.

Fig. 7.

As in Fig. 6, but with Niño-3.4 linearly removed for the positive and negative phases.

Removal of ENSO from the warm phase of the IOBM takes away much of the coherent signals seen in the left column of Fig. 6, because of the strong coherence between the positive phase of the IOBM and positive Niño-3.4 (left column, Fig. 7). In contrast, despite removal of ENSO from the cold phase of the IOBM, much of the anomalies associated with a negative IOBM phase remain (right column of Fig. 7). Two inferences emerge. First, an ENSO-incoherent positive IOBM phase does not have an impact on Australian rainfall that is proportionate to its amplitude. As discussed above, the impact from El Niño is not proportionate to its amplitude either, which is the reason why the positive IOBM does not have an effect on Australia rainfall; once the convection center moves away from Australia the amplitude of the positive IOBM becomes irrelevant.

Second, a La Niña–incoherent negative IOBM phase has a similar influence to that coherent with a La Niña; as a result, a similar tropical teleconnection operates extending the influence from northwest Australia (cf. right panels of Figs. 6 and 7), with consistent OLR and rainfall anomalies. This enhanced convection over northwest Australia during anomalously cool SST is likely due to a SST gradient produced in the region, as a result of the low SST coherence off the northwest coast of Australia (Fig. 6b). Further, a high pressure center over the Tasman Sea latitudes (Fig. 6f) emanating from northern Australia convection promotes onshore flows that are conducive to rainfall over the east coast, particularly the southern east coast.

To further confirm that the Rossby wave trains associated with the IOBM has little impact during El Niño, but enhances the impact of La Niña, we construct the 200-hPa height anomaly pattern associated with El Niño (Fig. 8a), El Niño events that are not coherent with the IOBM (Fig. 8c), La Niña (Fig. 8b), and La Niña events that are not coherent with the IOBM (Fig. 8d). In the El Niño cases, the PSA pattern is situated far away from Australia and the IOBM makes little difference to the PSA pattern (Figs. 8a,c). In the La Niña cases, the presence of the IOBM significantly enhances the strength of the high pressure center close to Australia (cf. Figs. 8b and 8d).

Fig. 8.

One standard deviation anomalies of 200-hPa geopotential height associated with (a) positive Niño-3.4 values, (b) negative Niño-3.4 values, (c) positive Niño-3.4 values with the IOBM linearly removed, and (d) negative Niño-3.4 values with the IOBM linearly removed.

Fig. 8.

One standard deviation anomalies of 200-hPa geopotential height associated with (a) positive Niño-3.4 values, (b) negative Niño-3.4 values, (c) positive Niño-3.4 values with the IOBM linearly removed, and (d) negative Niño-3.4 values with the IOBM linearly removed.

c. Asymmetry in the IOBM–northern Australia convection

The consequence of the asymmetric ENSO impact on Australia, with a greater influence during La Niña, would make the IOBM modulation on rainfall asymmetric. Given that the influence of the IOBM on the extratropics is conducted through wave trains emanating from northern Australia convective anomalies, a question arises as to whether the response of 200-hPa height anomalies over the Tasman Sea to the convective anomalies is itself asymmetric. This is examined in Fig. 9a, which shows that while a weak asymmetry exists, with a higher sensitivity to positive OLR anomalies, the response to both enhanced and decreased convection is statistically significant above the 99% confidence level. The higher sensitivity to positive OLR anomalies would have produced height anomalies over the Tasman Sea during the positive phase of the IOBM, which was not observed in our analysis.

Fig. 9.

Scatterplots of (a) 200-hPa geopotential height over the Tasman Sea region (47°–32°S, 150°–175°E), (b) IOBM, and (c) Niño-3.4 vs outgoing longwave radiation over northern Australia (25°–10°S, 125°–140°E). Linear regression coefficients (slopes), correlations, and p values of the linear fits conducted using all samples, only positive, and only negative values of the x axis are displayed.

Fig. 9.

Scatterplots of (a) 200-hPa geopotential height over the Tasman Sea region (47°–32°S, 150°–175°E), (b) IOBM, and (c) Niño-3.4 vs outgoing longwave radiation over northern Australia (25°–10°S, 125°–140°E). Linear regression coefficients (slopes), correlations, and p values of the linear fits conducted using all samples, only positive, and only negative values of the x axis are displayed.

We note, however, that the coherence of a positive IOBM with northern Australia convection is weaker than that of a negative IOBM (Fig. 9b) and also, the relationship between the convection anomalies and positive IOBM phases is not statistically significant (blue line, Fig. 9b). Thus, positive IOBM phases do not systematically induce convective anomalies, and El Niño itself generates little convection anomalies over northern Australia, because much of the convection anomalies are situated in the central equatorial Pacific. By contrast, enhanced convection over northern Australia displays a higher coherence with a negative phase of the IOBM and a La Niña phase (Figs. 9b,c). It is this symbiosis of the coherent northern Australia convection anomalies with negative phases of both the IOBM and La Niña, which ensures a greater effect of the negative IOBM.

6. ENSO–rainfall teleconnection through the SAM

As discussed in the introduction, in austral summer, ENSO is correlated with the SAM. L’Heureux and Thompson (2006) find that the forcing of ENSO on the SAM is through both thermally forced and eddy-driven zonal wind anomalies. A positive SAM is associated with a La Niña phase, with positive SLP anomalies embedded into the high SLP of the Tahiti center of the Southern Oscillation index, although the correlation is weak (Fig. 10a). Consistently, SLP anomalies associated with ENSO show a projection onto the midlatitude band (Fig. 10j), highlighting the link between ENSO and SAM that was shown in Fig. 2c. Reflecting this coherence with ENSO is a SAM–rainfall correlation pattern that is similar to that associated with ENSO (Figs. 10c,l). In particular, onshore flows associated with the midlatitude high SLP anomalies are conductive to the rainfall along the east coast of Australia.

Fig. 10.

(left to right) One standard deviation anomalies, based on the 1958–2011 period of SLP (hPa), SST (°C), and Australian rainfall (mm) associated with (top to bottom) the SAM index, the SAM index after removal of the ENSO coherent variance, Niño-3.4 after removal of variance coherent with the SAM, and Niño-3.4.

Fig. 10.

(left to right) One standard deviation anomalies, based on the 1958–2011 period of SLP (hPa), SST (°C), and Australian rainfall (mm) associated with (top to bottom) the SAM index, the SAM index after removal of the ENSO coherent variance, Niño-3.4 after removal of variance coherent with the SAM, and Niño-3.4.

However, removing ENSO from the SAM takes out much of the SAM-related rainfall anomalies, suggesting that the SAM-related anomalies mainly reflect its coherence with ENSO (Figs. 10a–c and 10d–f). Further, removing the SAM signals from ENSO generates no substantial change to ENSO anomaly and teleconnection patterns (Figs. 10g–i and 10j–l). We conclude that the SAM does not modulate the ENSO teleconnection in any major way.

7. Conclusions

In austral summer, ENSO covaries with the Indian Ocean through a strong coherence with the IOBM. ENSO also has a weak coherence with the SAM. The present study addresses how the IOBM and the SAM modulate the ENSO–rainfall teleconnection over Australia and whether the modulation is symmetric. These are examined in terms of the tropical and extratropical responses to ENSO. We show that the modulating effect of the SAM on ENSO is limited, mainly reflecting its relatively weak coherence with ENSO; in particular, the SAM does not modify the teleconnection pattern of ENSO. We confirm that the IOBM extends the ENSO-induced convection anomaly westward into northern Australia and the eastern Indian Ocean, allowing the ENSO tropical teleconnection to extend to inland Australia from the northwest, as discussed by previous studies (e.g., Taschetto et al. 2011).

We show that the IOBM also generates an equivalent-barotropic Rossby wave train, mostly through convection anomalies over northern Australia. The wave train shares an anomaly center over the Tasman Sea latitudes with the PSA pattern, which shifts the anomaly center of the PSA to within a greater proximity to Australia. This is in contrast to the suggestion from the modeling study of Taschetto et al. (2011), which found an anomaly center south of the Great Australian Bight.

Another contribution of the present study is the identification of a strong Australian rainfall modulation asymmetry between the positive and negative IOBM. During the negative phase, which tends to coincide with La Niña events, the modulating effect on rainfall is far greater than that during the positive phase of the IOBM, which tends to coincide with El Niño events. The La Niña teleconnection on Australian rainfall is stronger, providing a greater impact for the negative IOBM to modulate. La Niña–induced rainfall increase along northern and eastern Australia is proportionate to La Niña amplitude, whereas El Niño–induced rainfall reduction is not related to El Niño amplitude (Cai et al. 2010). This is because during La Niña, the convection center is situated over Australia and the impact increases with the amplitude of La Niña. Whereas during El Niño, the convection center moves eastward away from Australia; once the influence is remote from Australia a further increase in El Niño amplitude would not affect rainfall. This translates to the IOBM; so convective anomalies over northern Australia are more coherent with the negative phase of the IOBM than with the positive phase of the IOBM. As a consequence, there are more-coherent wave trains emanating from northern Australia during negative IOBM, which modulates the La Niña teleconnection.

Our results that Rossby-wave trains and their sources, which are associated with northern Australia, though consistent with previous studies (Jin and Hoskins 1995; DeWeaver and Nigam 2004), need to be tested through further modeling studies. In particular, the sensitivity of wave train anomaly centers to the location of convective anomalies in the western Pacific and eastern Indian Ocean needs to be tested. It would be valuable to examine if results of such sensitivity experiments reconcile the previous model results with what we presented here in terms of the pressure anomaly center over the Indian Ocean extratropics.

Acknowledgments

This work is supported by the Australia Climate Change Science Programme. We thank Evan Weller and Ariaan Purich for reviewing the paper before submission. We also thank three anonymous reviewers for their valuable comments.

REFERENCES

REFERENCES
Ashok
,
K.
,
Z.
Guan
, and
T.
Yamagata
,
2003
:
Influence of the Indian Ocean Dipole on the Australian winter rainfall
.
Geophys. Res. Lett.
,
30
,
1821
,
doi:10.1029/2003GL017926
.
Behera
,
S. K.
, and
T.
Yamagata
,
2001
:
Subtropical SST dipole events in the southern Indian Ocean
.
Geophys. Res. Lett.
,
28
,
327
330
.
Behera
,
S. K.
,
J.-J.
Luo
,
S.
Masson
,
P.
Delecluse
,
S.
Gualdi
,
A.
Navarra
, and
T.
Yamagata
,
2005
:
Paramount impact of the Indian Ocean Dipole on the East African short rain: A CGCM study
.
J. Climate
,
18
,
4514
4530
.
Black
,
E.
,
J. M.
Slingo
, and
K. R.
Sperber
,
2003
:
An observational study of the relationship between excessively strong short rains in coastal East Africa and Indian Ocean SST
.
Mon. Wea. Rev.
,
131
,
74
94
.
Cai
,
W.
, and
P.
van Rensch
,
2012
:
The 2011 southeast Queensland extreme summer rainfall: A confirmation of a negative Pacific Decadal Oscillation phase?
Geophys. Res. Lett.
,
39
,
L08702
,
doi:10.1029/2011GL050820
.
Cai
,
W.
,
H. H.
Hendon
, and
G. A.
Meyers
,
2005
:
Indian Ocean dipole-like variability in the CSIRO Mark 3 coupled climate model
.
J. Climate
,
18
,
1449
1468
.
Cai
,
W.
,
T.
Cowan
, and
M.
Raupach
,
2009a
:
Positive Indian Ocean Dipole events precondition southeast Australia bushfires
.
Geophys. Res. Lett.
,
36
,
L19710
,
doi:10.1029/2009GL039902
.
Cai
,
W.
,
T.
Cowan
, and
A.
Sullivan
,
2009b
:
Recent unprecedented skewness towards positive Indian Ocean Dipole occurrences and its impact on Australian rainfall
.
Geophys. Res. Lett.
,
36
,
L11705
,
doi:10.1029/2009GL037604
.
Cai
,
W.
,
P.
van Rensch
,
T.
Cowan
, and
A.
Sullivan
,
2010
:
Asymmetry in ENSO teleconnection with regional rainfall, its multidecadal variability, and impact
.
J. Climate
,
23
,
4944
4955
.
Cai
,
W.
,
P.
van Rensch
,
T.
Cowan
, and
H. H.
Hendon
,
2011
:
Teleconnection pathways of ENSO and the IOD and the mechanisms for impacts on Australian rainfall
.
J. Climate
,
24
,
3910
3923
.
Cai
,
W.
,
P.
van Rensch
,
T.
Cowan
, and
H. H.
Hendon
,
2012a
:
An asymmetry in the IOD and ENSO teleconnection pathway and its impact on Australian climate
.
J. Climate
,
25
,
6318
6329
.
Cai
,
W.
, and
Coauthors
,
2012b
:
More extreme swings of the South Pacific Convergence Zone due to greenhouse warming
.
Nature
,
488
,
365
369
,
doi:10.1038/nature11358
.
Chambers
,
D. P.
,
B. D.
Tarpley
, and
R. H.
Stewart
,
1999
:
Anomalous warming in the Indian Ocean coincident with El Niño
.
J. Geophys. Res.
,
104
(
C2
),
3035
3047
.
DeWeaver
,
E.
, and
S.
Nigam
,
2004
:
On the forcing of ENSO teleconnections by anomalous heating and cooling
.
J. Climate
,
17
,
3225
3235
.
Du
,
Y.
,
S.-P.
Xie
,
G.
Huang
, and
K.
Hu
,
2009
:
Role of air–sea interaction in the long persistence of El Niño–induced North Indian Ocean warming
.
J. Climate
,
22
,
2023
2038
.
Fauchereau
,
N.
,
S.
Trzaska
,
Y.
Richard
,
P.
Roucou
, and
P.
Camberlin
,
2003
:
Sea-surface temperature co-variability in the Southern Atlantic and Indian Oceans and its connections with the atmospheric circulation in the Southern Hemisphere
.
Int. J. Climatol.
,
23
,
663
677
,
doi:10.1002/joc.905
.
Folland
,
C. K.
,
J. A.
Renwick
,
M. J.
Salinger
, and
A. B.
Mullan
,
2002
:
Relative influence of the interdecadal Pacific oscillation and ENSO on the South Pacific Convergence Zone
.
Geophys. Res. Lett.
,
29
,
1643
,
doi:10.1029/2001GL014201
.
Gill
,
A. E.
,
1980
:
Some simple solutions for heat-induced tropical circulation
.
Quart. J. Roy. Meteor. Soc.
,
106
,
447
462
.
Hoskins
,
B. J.
, and
D. J.
Karoly
,
1981
:
The steady linear response of a spherical atmosphere to thermal and orographic forcing
.
J. Atmos. Sci.
,
38
,
1179
1196
.
Jin
,
F.-F.
, and
B. J.
Hoskins
,
1995
:
The direct response to tropical heating in a baroclinic atmosphere
.
J. Atmos. Sci.
,
52
,
307
319
.
Jones
,
D. A.
,
W.
Wang
, and
R.
Fawcett
,
2009
:
High-quality spatial climate data-sets for Australia
.
Aust. Meteor. Oceanogr. J.
,
58
,
233
248
.
Kalnay
,
E.
, and
Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
.
Klein
,
S. A.
,
B. J.
Soden
, and
N.-C.
Lau
,
1999
:
Remote sea surface variations during ENSO: Evidence for a tropical atmospheric bridge
.
J. Climate
,
12
,
917
932
.
L’Heureux
,
M. L.
, and
D. W. J.
Thompson
,
2006
:
Observed relationships between the El Niño–Southern Oscillation and the extratropical zonal-mean circulation
.
J. Climate
,
19
,
276
287
.
Marshall
,
G. J.
,
2003
:
Trends in the southern annular mode from observations and reanalyses
.
J. Climate
,
16
,
4134
4143
.
McBride
,
J. L.
, and
N.
Nicholls
,
1983
:
Seasonal relationships between Australian rainfall and the Southern Oscillation
.
Mon. Wea. Rev.
,
111
,
1998
2004
.
Meyers
,
G. A.
,
P. C.
McIntosh
,
L.
Pigot
, and
M. J.
Pook
,
2007
:
The years of El Niño, La Niña, and interactions with the tropical Indian Ocean
.
J. Climate
,
20
,
2872
2880
.
Nicholls
,
N.
,
1985
:
Towards the prediction of major Australian droughts
.
Aust. Meteor. Mag.
,
33
,
161
166
.
North
,
G. R.
,
T. L.
Bell
,
R. F.
Cahalan
, and
F. J.
Moeng
,
1982
:
Sampling errors in the estimation of empirical orthogonal functions
.
Mon. Wea. Rev.
,
110
,
699
706
.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
,
doi:10.1029/2002JD002670
.
Reason
,
C. J. C.
,
2001
:
Subtropical Indian Ocean SST dipole events and southern African rainfall
.
Geophys. Res. Lett.
,
28
,
2225
2227
.
Saji
,
N. H.
, and
T.
Yamagata
,
2003
:
Possible impacts of Indian Ocean Dipole mode events on global climate
.
Climate Res.
,
25
,
151
169
.
Taschetto
,
A. S.
,
A.
Sen Gupta
,
H. H.
Hendon
,
C. C.
Ummenhofer
, and
M. H.
England
,
2011
:
The contribution of Indian Ocean sea surface temperature anomalies on Australian summer rainfall during El Niño events
.
J. Climate
,
24
,
3734
3747
.
Venegas
,
S.
,
L. A.
Mysak
, and
D. N.
Straub
,
1997
:
Atmosphere–ocean coupled variability in the South Atlantic
.
J. Climate
,
10
,
2904
2920
.
Vincent
,
E. M.
,
M.
Lengaigne
,
C. E.
Menkes
,
N. C.
Jourdain
,
P.
Marchesiello
, and
G.
Madec
,
2011
:
Interannual variability of the South Pacific Convergence Zone and implications for tropical cyclone genesis
.
Climate Dyn.
,
36
,
1881
1896
,
doi:10.1007/s00382-009-0716-3
.
Xie
,
S.-P.
,
H.
Annamalai
,
F. A.
Schott
, and
J. P.
McCreary
,
2002
:
Structure and mechanisms of South Indian Ocean climate variability
.
J. Climate
,
15
,
864
878
.
Xie
,
S.-P.
,
K.
Hu
,
J.
Hafner
,
H.
Tokinaga
,
Y.
Du
,
G.
Huang
, and
T.
Sampe
,
2009
:
Indian Ocean capacitor effect on Indo–western Pacific climate during the summer following El Niño
.
J. Climate
,
22
,
730
747
.
Zubair
,
L.
,
S. A.
Rao
, and
T.
Yamagata
,
2003
:
Modulation of Sri Lankan Maha rainfall by the Indian Ocean dipole
.
Geophys. Res. Lett.
,
30
,
1063
,
doi:10.1029/2002GL015639
.