• 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.

    • Search Google Scholar
    • Export Citation
  • Cai, W., and T. Cowan, 2008: Dynamics of late autumn rainfall reduction over southeastern Australia. Geophys. Res. Lett., 35, L09708, doi:10.1029/2008GL033727.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Cai, W., A. Pan, D. Roemmich, T. Cowan, and X. Guo, 2009c: Argo profiles a rare occurrence of three consecutive positive Indian Ocean Dipole events, 2006–2008. Geophys. Res. Lett., 36, L08701, doi:10.1029/2008GL037038.

    • Search Google Scholar
    • Export Citation
  • Cai, W., A. Sullivan, T. Cowan, J. Ribbe, and G. Shi, 2011: Simulation of the Indian Ocean Dipole: A relevant criterion for selecting models for climate projections. Geophys. Res. Lett., 38, L03704, doi:10.1029/2010GL046242.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and S. Nigam, 2004: On the forcing of ENSO teleconnections by anomalous heating and cooling. J. Climate, 17, 32253235.

  • Fischer, A. P., P. Terray, E. Guilyardi, S. Gualdi, and P. Delecluse, 2005: Two independent triggers for the Indian Ocean dipole/zonal mode in a coupled GCM. J. Climate, 18, 34283449.

    • Search Google Scholar
    • Export Citation
  • Frederiksen, C. S., and J. S. Frederiksen, 1996: A theoretical model of Australian northwest cloudband disturbances and Southern Hemisphere storm tracks: The role of SST anomalies. J. Atmos. Sci., 53, 14101432.

    • Search Google Scholar
    • Export Citation
  • Ghil, M., and K. C. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part II: Southern Hemisphere. J. Atmos. Sci., 48, 780790.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447462.

  • Hendon, H. H., D. W. J. Thompson, and M. C. Wheeler, 2007: Australian rainfall and surface temperature variations associated with the Southern Hemisphere Annular Mode. J. Climate, 20, 24522467.

    • Search Google Scholar
    • Export Citation
  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109, 813829.

    • Search Google Scholar
    • Export Citation
  • 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, 11791196.

    • Search Google Scholar
    • Export Citation
  • Jin, F., and B. J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52, 307319.

  • Jones, D., and G. Weymouth, 1997: An Australian monthly rainfall dataset. Tech. Rep. 70, Bureau of Meteorology, 19 pp.

  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events. J. Climate, 2, 12391252.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 12751277.

    • Search Google Scholar
    • Export Citation
  • 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, 28722880.

    • Search Google Scholar
    • Export Citation
  • Murphy, B., and B. Timbal, 2008: A review of recent climate variability and climate change in south-eastern Australia. Int. J. Climatol., 28, 859879.

    • Search Google Scholar
    • Export Citation
  • Nicholls, N., 1985: Towards the prediction of major Australian droughts. Aust. Meteor. Mag., 33, 161166.

  • Nicholls, N., 1989: Sea surface temperatures and Australian winter rainfall. J. Climate, 2, 965973.

  • Nicholls, N., 2009: Local and remote causes of the southern Australian autumn–winter rainfall decline, 1958–2007. Climate Dyn., 34, 835845, doi:10.1007/s00382-009-0527-6.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and J. Kim, 1997: The relationship of the El Niño–Southern Oscillation to African rainfall. Int. J. Climatol., 17, 117135.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Risbey, J. S., M. J. Pook, P. C. McIntosh, C. C. Ummenhofer, and G. Meyers, 2009: Characteristics and variability of synoptic features associated with cool season rainfall in southeastern Australia. Int. J. Climatol., 29, 15951613.

    • Search Google Scholar
    • Export Citation
  • Saji, N. H., and T. Yamagata, 2003: Possible impacts of Indian Ocean Dipole mode events on global climate. Climate Res., 25, 151169.

  • Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360363.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45, 12281251.

    • Search Google Scholar
    • Export Citation
  • Tapp, R. G., and S. L. Barrell, 1984: The north–west Australian cloud band: Climatology, characteristics and factors associated with development. Int. J. Climatol., 4, 411424.

    • Search Google Scholar
    • Export Citation
  • Ummenhofer, C. C., A. S. Gupta, M. J. Pook, and M. H. England, 2008: Anomalous rainfall over southwest Western Australia forced by Indian Ocean sea surface temperatures. J. Climate, 21, 51135134.

    • Search Google Scholar
    • Export Citation
  • Ummenhofer, C. C., A. S. T. A. Sen Gupta, and M. H. England, 2009: Modulation of Australian precipitation by meridional gradients in east Indian Ocean sea surface temperature. J. Climate, 22, 55975610.

    • Search Google Scholar
    • Export Citation
  • Walker, G. T., 1923: Correlations in seasona variations in weather, VIII. Mem. India Meteor. Dept., 24, 75131.

  • Watterson, I. G., 1987: The effect of the Hadley Circulation on the meridional propagation of stationary waves. Quart. J. Roy. Meteor. Soc., 113, 779813.

    • Search Google Scholar
    • Export Citation
  • Watterson, I. G., 2010: Relationships between southeastern Australian rainfall and sea surface temperatures examined using a climate model. J. Geophys. Res., 115, D10108, doi:10.1029/2009JD012120.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., A. M. Moore, J. P. Loschnigg, and R. R. Leben, 1999: Coupled ocean-atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356360.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., and H. H. Hendon, 2009: Representation and prediction of the Indian Ocean Dipole in the POAMA seasonal forecast model. Quart. J. Roy. Meteor. Soc., 135, 337352.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Correlations between Niño-3.4, the DMI, and the individual poles of the DMI (IODE and IODW) for austral (a) winter and (c) spring. Red correlations indicate significant correlation at the 5% level. Correlations between Niño-3.4 and the DMI with OLR in the IODE and IODW regions are shown for (b) winter and (d) spring.

  • View in gallery

    Regressions of (left) SST (°C), (middle) OLR (W m−2), and (right) 500-hPa level vertical velocity (Pa s−1) onto (from top to bottom) Niño-3.4, partial Niño-3.4, partial DMI, and DMI for the austral winter season (JJA). Anomalies are shown for a one standard deviation anomaly of the predictor in each case. Bold green contours encompass the statistically significant regression coefficients at the 95% confident interval assuming 28 degrees of freedom.

  • View in gallery

    Regression of (left) MSLP (hPa), (middle) Z200 (m), and (right) Australian rainfall (mm) onto (from top to bottom) Niño-3.4, partial Niño-3.4, partial DMI, and DMI for the austral winter season (JJA). The plotting convention is as in Fig. 2. Contours of correlation coefficient of magnitude 0.6 and 0.8 are shown in MSLP and Z200 regression fields.

  • View in gallery

    Regression of var500 (m2) onto (a) Niño-3.4, (b) partial Niño-3.4, (c) partial DMI, and (d) DMI for austral winter (JJA). The plotting convention is as in Fig. 2.

  • View in gallery

    As in Fig. 2 but for austral spring (SON).

  • View in gallery

    As in Fig. 3 but for austral spring (SON).

  • View in gallery

    As in Fig. 4 but for austral spring (SON).

  • View in gallery

    Partial regression of Australian rainfall (mm) onto Niño-3.4 after the effects of OLR in the IODE region have been removed for the austral spring season (SON). The plotting convention is as in Fig. 2.

  • View in gallery

    Schematic illustration of the typical wave trains associated with the (a) IOD and (b) ENSO for JJA and (c) for the IOD and ENSO together for SON. This description is for positive phases of the IOD and El Niño. Typically opposite patterns occur during negative IOD and La Niña. Shaded blue (red) areas indicate regions of increased (decreased) tropical convection. Blue (red) contours indicate anomalously low (high) upper-level heights. The dashed lines trace the prominent wave trains mentioned in the text: gray for the EIO wave train, green for the WIO wave train, and orange for the EqA wave train, PNA and PSA wave trains not marked but visible in (b) and (c).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 2188 1135 78
PDF Downloads 1809 985 55

Teleconnection Pathways of ENSO and the IOD and the Mechanisms for Impacts on Australian Rainfall

View More View Less
  • 1 CSIRO Wealth from Oceans National Research Flagship, and CSIRO Water for a Healthy Country Flagship, Aspendale, Victoria, Australia
  • | 2 Centre for Australian Weather and Climate Research, Bureau of Meteorology, Melbourne, Australia
Full access

Abstract

Impacts of El Niño–Southern Oscillation (ENSO) and the Indian Ocean dipole (IOD) on Australian rainfall are diagnosed from the perspective of tropical and extratropical teleconnections triggered by tropical sea surface temperature (SST) variations. The tropical teleconnection is understood as the equatorially trapped, deep baroclinic response to the diabatic (convective) heating anomalies induced by the tropical SST anomalies. These diabatic heating anomalies also excite equivalent barotropic Rossby wave trains that propagate into the extratropics. The main direct tropical teleconnection during ENSO is the Southern Oscillation (SO), whose impact on Australian rainfall is argued to be mainly confined to near-tropical portions of eastern Australia. Rainfall is suppressed during El Niño because near-tropical eastern Australia directly experiences subsidence and higher surface pressure associated with the western pole of the SO. Impacts on extratropical Australian rainfall during El Niño are argued to stem primarily from the Rossby wave trains forced by convective variations in the Indian Ocean, for which the IOD is a primary source of variability. These equivalent-barotropic Rossby wave trains emanating from the Indian Ocean induce changes to the midlatitude westerlies across southern Australia, thereby affecting rainfall through changes in mean-state baroclinicity, west–east steering, and possibly orographic effects. Although the IOD does not mature until austral spring, its impact on Australian rainfall during winter is also ascribed to this mechanism. Because ENSO is largely unrelated to the IOD during austral winter, there is limited impact of ENSO on rainfall across southern latitudes of Australia in winter. A strong impact of ENSO on southern Australia rainfall in spring is ascribed to the strong covariation of ENSO and the IOD in this season. Implications of this pathway from the tropical Indian Ocean for impacts of both the IOD and ENSO on southern Australian climate are discussed with regard to the ability to make seasonal climate predictions and with regard to the role of trends in tropical SST for driving trends in Australian climate.

Corresponding author address: Wenju Cai, CSIRO Marine and Atmospheric Research, PMB 1, Aspendale, VIC 3195, Australia. E-mail: wenju.cai@csiro.au

Abstract

Impacts of El Niño–Southern Oscillation (ENSO) and the Indian Ocean dipole (IOD) on Australian rainfall are diagnosed from the perspective of tropical and extratropical teleconnections triggered by tropical sea surface temperature (SST) variations. The tropical teleconnection is understood as the equatorially trapped, deep baroclinic response to the diabatic (convective) heating anomalies induced by the tropical SST anomalies. These diabatic heating anomalies also excite equivalent barotropic Rossby wave trains that propagate into the extratropics. The main direct tropical teleconnection during ENSO is the Southern Oscillation (SO), whose impact on Australian rainfall is argued to be mainly confined to near-tropical portions of eastern Australia. Rainfall is suppressed during El Niño because near-tropical eastern Australia directly experiences subsidence and higher surface pressure associated with the western pole of the SO. Impacts on extratropical Australian rainfall during El Niño are argued to stem primarily from the Rossby wave trains forced by convective variations in the Indian Ocean, for which the IOD is a primary source of variability. These equivalent-barotropic Rossby wave trains emanating from the Indian Ocean induce changes to the midlatitude westerlies across southern Australia, thereby affecting rainfall through changes in mean-state baroclinicity, west–east steering, and possibly orographic effects. Although the IOD does not mature until austral spring, its impact on Australian rainfall during winter is also ascribed to this mechanism. Because ENSO is largely unrelated to the IOD during austral winter, there is limited impact of ENSO on rainfall across southern latitudes of Australia in winter. A strong impact of ENSO on southern Australia rainfall in spring is ascribed to the strong covariation of ENSO and the IOD in this season. Implications of this pathway from the tropical Indian Ocean for impacts of both the IOD and ENSO on southern Australian climate are discussed with regard to the ability to make seasonal climate predictions and with regard to the role of trends in tropical SST for driving trends in Australian climate.

Corresponding author address: Wenju Cai, CSIRO Marine and Atmospheric Research, PMB 1, Aspendale, VIC 3195, Australia. E-mail: wenju.cai@csiro.au

1. Introduction

This study is motivated by the need to better understand how El Niño–Southern Oscillation (ENSO) affects Australian climate. At first thought, this would appear to be a solved problem especially because the impact of ENSO on Australian climate, in particular the tendency for reduced rainfall/drought to accompany El Niño, has been known for many decades (e.g., Walker 1923); this impact is the premise of extended-range prediction of Australian climate (e.g., Nicholls 1985). However, it is not obvious which sea surface temperature (SST) variations in the tropics during ENSO are the sources of Australian climate variability. For instance, SST in the tropical Indian Ocean covaries with that in the tropical Pacific during ENSO and impacts on Australian rainfall have been related to SST variations in the tropical Indian Ocean (e.g., Nicholls 1989). Furthermore, the Indian Ocean dipole (IOD) (e.g., Saji et al. 1999; Webster et al. 1999) is recognized now to be a key component of the SST variability in the tropical Indian Ocean, contributing to rainfall variability that is both dependent and independent of ENSO. For instance, Risbey et al. (2009) showed that cool-season Australian rainfall (June–November) is typically reduced throughout all of eastern and extreme southern parts of Australia during El Niño but, when the covariance with the IOD is removed, the impacts across southern Australia disappear. Similarly, impacts of the IOD are felt throughout eastern and southern Australia, but when the covariance with ENSO is removed, the impacts of the IOD are only felt across the south. Meyers et al. (2007) reach a similar conclusion based on composite analysis for ENSO and IOD years. These results imply both that ENSO’s impacts on Australian climate are strengthened in the presence of the IOD and that ENSO and the IOD affect northeastern and southern Australian rainfall via distinct mechanisms: one has its roots in the coupled variability of the Pacific Ocean [as expressed by development of the Southern Oscillation (SO)] and mainly affects lower latitudes of eastern Australia. The other has its roots in the Indian Ocean and mainly impacts higher latitudes and more western portions of Australia (see also Nicholls 1989). This suggestion of separate, but interrelated, pathways needs clarification.

Although the motivation here is to better understand the underlying mechanisms for the transmission of ENSO and the IOD to Australian climate, this study will more broadly address how the effects of ENSO and the IOD are transmitted into the Southern Hemisphere. The perspective here is to better understand how variations of tropical SST during ENSO and the IOD drive remote atmospheric responses. The fundamental mechanisms for these responses are, of course, well established. For instance, in the tropics, the direct atmospheric response to El Niño is the SO: an anomalous, hemispheric, zonal dipole in surface pressure with its east (west) pole near 120°W (120°E). During El Niño the pressure in the east (west) pole is anomalously low (high) in conjunction with warm (cool) SST anomalies and enhanced (reduced) rainfall in the central Pacific (the Maritime Continent); the opposite occurs during La Niña.

The development of the SO, at least its tropical component, is well understood as the thermally direct response of the tropical atmosphere to the diabatic heating associated with the anomalous zonal dipole in rainfall (e.g., Gill 1980): lower-tropospheric cyclones (low surface pressure) straddle the positive heating anomalies in the eastern Pacific and anticyclones (high surface pressure) straddle the negative heating anomaly in the Maritime Continent/western Pacific. Because the heating anomalies extend throughout the depth of the troposphere and are primarily balanced by adiabatic motion (upward in regions of anomalous heating and downward in regions of anomalous cooling), the response in the upper troposphere is of opposite sign so that anticyclones overlie the cyclones in the east and cyclones overlie the anticyclones in the west. As a result of the variation in the Coriolis parameter with latitude, the deep baroclinic component of the SO is trapped to within about 25° latitude of the equator.

In addition to the deep baroclinic response that is equatorially trapped, the zonal dipole in anomalous rainfall during ENSO also acts as a Rossby wave source, whereby equivalent barotropic Rossby wave trains are excited by the regions of upper-level divergence (convergence) that occur over the regions of enhanced (suppressed) rainfall. These Rossby wave trains “teleconnect” the effects of ENSO into higher latitudes of both hemispheres (e.g., Hoskins and Karoly 1981). In the Northern Hemisphere, the primary Rossby wave train response to the anomalous rainfall in the central Pacific during ENSO is referred to as the Pacific–North American pattern (PNA) (Horel and Wallace 1981). The PNA pattern modulates the extratropical jet streams and storm tracks in the northeast Pacific and across North America, thereby altering the climate in these regions during ENSO. The equivalent pattern extending into the central and eastern south Pacific and arching toward South America is referred to as the Pacific–South American (PSA) pattern (e.g., Karoly 1989; Ghil and Mo 1991).

Although ENSO is also associated with upper-tropospheric divergence anomalies over the eastern Indian Ocean/far western Pacific that could potentially act as Rossby wave sources (e.g., DeWeaver and Nigam 2004), the PNA and PSA patterns are dominant (and have received the most attention) because the upper-level divergence anomalies during ENSO in the central Pacific occur in regions where the time-mean upper-level westerly winds, required for Rossby wave propagation, extend nearest to the equator. Sardeshmukh and Hoskins (1988) showed, however, that divergence anomalies in regions of upper-level easterlies, such as those that occur across the Maritime Continent, can still act as effective Rossby wave sources because of advection of anomalous vorticity by the mean divergent wind. Additionally, a mean meridional wind occurring concurrently with upper-level easterlies has been shown to transport anomalous vorticity in the direction of the wind (Watterson 1987). Jin and Hoskins (1995) demonstrated, for instance, that an equatorial divergence anomaly over the Maritime Continent imposed on the December–February mean flow will excite Rossby wave trains that propagate into both hemispheres.

Because the Australian continent extends into the tropics, it can be assumed that near-tropical portions of eastern Australia directly feel impacts of the baroclinic component of the SO (e.g., dry during El Niño when local surface pressure is high and air is subsiding). However, the baroclinic component of the SO cannot directly account for climate anomalies across southern Australia because it is equatorially trapped. Furthermore, the PSA pattern cannot account for the anomalies across the higher latitudes of Australia because it occurs well to the east of Australia. This points to the tropical rainfall anomalies in the vicinity of the Maritime Continent as the probable source of Rossby wave trains that drive climate variations over extratropical Australia during ENSO and the IOD. This mechanism has received little attention for explaining ENSO variations in higher latitudes of Australia, although this pathway during the IOD is appreciated (e.g., Saji and Yamagata 2003; Cai et al. 2009b).

Although there has been an establishment of the statistical link between the IOD and extratropical Australian rainfall (e.g., Nicholls 1989; Ashok et al. 2003; Cai et al. 2009b; Risbey et al. 2009) and recognition of a Rossby wave train emanating from the eastern Indian Ocean (e.g., Saji and Yamagata 2003), many questions remain. These include the relative and independent roles of ENSO and the IOD for modulating tropical rainfall to the north of Australia, which may act to excite Rossby wave trains, and the mechanism by which these equivalent barotropic wave trains then affect higher-latitude Australian rainfall (e.g., Ummenhofer et al. 2008). There is also the issue of the relative roles of the SST anomalies in the western and eastern poles of the IOD and whether the SST anomalies in the western pole matter for generating of a remote response (e.g., Nicholls 2009).

To address these issues in the present study, we will diagnose the tropical rainfall and associated tropical and extratropical circulation anomalies that develop in association with tropical SST anomalies during ENSO and the IOD. We focus on austral winter and spring, which are the cool seasons in Australia during which the impacts of ENSO and IOD on rainfall are greatest. Through use of regression and partial regression analysis we will confirm that the direct impacts of ENSO in Australia due to swings in the SO are confined to near-tropical eastern portions of the continent. In contrast, we will show that extratropical impacts over Australia from both ENSO and the IOD stem from Rossby wave trains that emanate from the tropical Indian Ocean. This pathway from the Indian Ocean through which not only the IOD but also ENSO impacts extratropical Australia is not fully appreciated but, for instance, has great implications for the ability to predict Australian climate even during ENSO and for assessing climate impacts stemming from trends in tropical SST. We detail our methods and datasets in section 2 and describe results for austral winter and austral spring in sections 3 and 4, respectively. The discussion and conclusions are provided in section 5.

2. Data and methods

a. Data

We focus on the period 1979–2008 when more reliable observations, particularly from satellites, are available. We divide the Australian cool season into austral winter (June–August) and spring (September–November), which are the seasons when ENSO and the IOD are known to have a strong influence on Australian climate. Although the IOD typically matures in austral spring when it is strongly correlated with ENSO, it develops in winter at which time it is much less correlated to ENSO (see section 2c). ENSO tends not to peak until austral summer, but it is often well developed by austral winter and spring. Hence, independent and dependent impacts of IOD and ENSO are felt in both seasons.

Tropical SST variations associated with ENSO and the IOD are monitored using the monthly Hadley Centre global sea Ice and SST analyses (HadISST) (Rayner et al. 2003). Associated variations in deep tropical convection are diagnosed with monthly mean outgoing longwave radiation (OLR) (Liebmann and Smith 1996). Upper-tropospheric divergence anomalies resulting from tropical convection variations, which may act as a Rossby wave source, are inferred with the vertical velocity at 500 hPa from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalyses (Kalnay et al. 1996). Resulting circulation anomalies are diagnosed using monthly mean sea level pressure (MSLP) and geopotential heights at 200 hPa (Z200) from the NCEP–NCAR reanalyses. Variations of synoptic-scale weather systems that are the primary source of extratropical rainfall are assessed using 2–7-day bandpass-filtered 500-hPa geopotential heights. After filtering, daily values at each grid point are squared and averaged monthly to obtain synoptic height variance (var500). Rainfall across Australia is diagnosed using a monthly gridded analyses (0.25° grid) based on available station observations (Jones and Weymouth 1997). Throughout this study, seasonal anomalies (3-month means) are constructed from a monthly 1979–2008 climatology. All time series are linearly detrended in order to focus on interannual variations.

ENSO is monitored using the Niño-3.4 SST index, which is the average SST anomalies over 5°S–5°N, 170°–120°W. The IOD is monitored with the dipole mode index (DMI) (Saji et al. 1999), which is the difference of the area mean SST anomalies in the IODW region (10°S–10°N, 50°–70°E) and the IODE region (10°S–0°, 90°–110°E.)

b. Regression technique

Our basic analysis technique is regression (or equivalently correlation) and partial regression (partial correlation). Partial regression is employed to isolate impacts of ENSO and the IOD that are independent of each other. Partial regression involves computation of the linear dependence of a predictand upon a predictor after the linear relationship with a second predictor has been removed from both the predictand and predictor. We will refer to the Niño-3.4 index (one key predictor) after the linear relation with the IOD index (the other key predictor) has been removed as Niño-3.4|DMI. Similarly the DMI index after the Niño-3.4 dependency has been removed is referred to as DMI|Niño3.4.

An anomalous circulation pattern associated with ENSO after removing the effects of the IOD may be described by a map of the partial correlation, C(x, y)Niño-3.4|DMI, which is calculated by removing the IOD-coherent variance from gridpoint anomalies of a field A(x, y, t) to obtain a residual field A(x, y, t)|DMI, and then correlating A(x, y, t)|DMI with Niño-3.4|DMI. Likewise, a map of partial regression [R(x, y)Niño-3.4|DMI] is constructed by regressing A(x, y, t)|DMI onto Niño-3.4|DMI. It is important to note that removing the influence of the IOD in this manner also removes the ENSO signal that is coherent with the IOD (and, similarly, for removing ENSO from the IOD). For display purposes, we scale the regression anomalies by a one standard deviation anomaly of the predictor in each case. Such a scaling facilitates a comparison of the size of anomalies associated with typical variations of the predictor indices. A comparison of unscaled regression coefficients often does not bring out such contrasts in the size. Significance of the regression coefficients are estimated based on significance of the associated correlation coefficients, which is when p < 0.05 with 28 degrees of freedom (i.e., the absolute value of the correlation coefficient is greater than 0.361).

c. Interdependence of predictors

Prior to exploring the circulation patterns and wave train patterns associated with the SST variations due to ENSO and the IOD, we first review the interdependence of our two main predictors, Niño-3.4 and the DMI. We do this by computing correlations between the two indices seasonally for austral winter (Fig. 1a) and austral spring (Fig. 1c). We also consider correlations with the individual poles of the DMI (IODE and IODW) to understand the relative roles of the SST variations in the western and eastern portions of the Indian Ocean to variability of the IOD and its relationship with ENSO. These correlations show that 1) Niño-3.4 is only moderately correlated with the DMI in June–August (JJA), 0.41, but strongly correlated in September–November (SON), 0.72; 2) Niño-3.4 is more strongly correlated with the DMI than it is with either of the individual poles of the IOD; 3) the DMI accounts for more variability in either of the IOD poles than does Niño-3.4; and 4), although the two poles of the IOD are not significantly anticorrelated in either season, the DMI accounts for a significant fraction of the variability in each pole, especially in SON and especially in the eastern pole.

Fig. 1.
Fig. 1.

Correlations between Niño-3.4, the DMI, and the individual poles of the DMI (IODE and IODW) for austral (a) winter and (c) spring. Red correlations indicate significant correlation at the 5% level. Correlations between Niño-3.4 and the DMI with OLR in the IODE and IODW regions are shown for (b) winter and (d) spring.

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

3. Winter anomalies

We begin by examining the regression patterns of anomalies of tropical SST, OLR (tropical convection), and vertical velocity at 500 hPa associated with Niño-3.4, Niño-3.4|DMI, DMI|Niño-3.4, and DMI for austral winter (Fig. 2). We immediately observe that the midtropospheric vertical velocity anomalies in the tropics have a similar spatial structure as the OLR anomalies, confirming that negative OLR is associated with upward motion in the midtroposphere (implying upper-troposphere divergence) and opposite sign for positive OLR.

Fig. 2.
Fig. 2.

Regressions of (left) SST (°C), (middle) OLR (W m−2), and (right) 500-hPa level vertical velocity (Pa s−1) onto (from top to bottom) Niño-3.4, partial Niño-3.4, partial DMI, and DMI for the austral winter season (JJA). Anomalies are shown for a one standard deviation anomaly of the predictor in each case. Bold green contours encompass the statistically significant regression coefficients at the 95% confident interval assuming 28 degrees of freedom.

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

The tropical SST and convection/divergence anomalies during austral winter associated with Niño-3.4 (Fig. 2) appear as a zonal dipole between the east Pacific and west Pacific/Maritime Continent/far eastern Indian Ocean region. The anomalies associated with the IOD are distinct from those associated with ENSO and are primarily associated with a zonal dipole in the equatorial eastern and western Indian Ocean and with very limited anomalies in the eastern Pacific. The partial regression patterns in each case are nearly identical to the full regression, reflecting the relatively weak relationship between Niño-3.4 and the DMI in winter (Fig. 1a). Furthermore, variations in the DMI are dominated by SST variations in the IODE where anomalous convection, as measured by anomalous OLR, is dominated by local SST anomalies (Fig. 1b).

With respect to Australian climate, downward motion over cool SST in the Maritime Continent region during El Niño extends poleward to cover northeastern portions of Australia (Figs. 2c,f). Importantly, this downward motion over northeastern Australia is not evident in the regression onto the DMI. Rather, the IOD is associated with a “tail” of high OLR, and downward motion extends southeastward from the equatorial eastern Indian Ocean into northwest Australia. This signature in OLR, which presumably stems primarily from variations of high cirrus clouds, has been associated with the occurrence of “northwest cloud bands” (Tapp and Barrell 1984), which are more frequent when the eastern tropical Indian Ocean is warm, as during negative IOD years (Frederiksen and Frederiksen 1996; Ummenhofer et al. 2008). The association of northwest cloud bands with rainfall variations across southern portions of the continent stemming from IOD variability (discussed below but evident in the weak, but significant, positive OLR anomalies across southern Australia in Figs. 2h,k) will be argued to be indirect (i.e., the northwest cloud bands are not just acting as moisture conduits from the tropics to the extratropics).

The circulation anomalies associated with ENSO and the IOD are assessed with regressions of MSLP and Z200 onto Niño-3.4 and the DMI (Fig. 3). As for SST/convection, the total regressions and partial regressions in each case are very similar owing to the relative independence of ENSO and the IOD in winter. We will, however, subsequently point out some subtle but important differences in the partial regressions.

Fig. 3.
Fig. 3.

Regression of (left) MSLP (hPa), (middle) Z200 (m), and (right) Australian rainfall (mm) onto (from top to bottom) Niño-3.4, partial Niño-3.4, partial DMI, and DMI for the austral winter season (JJA). The plotting convention is as in Fig. 2. Contours of correlation coefficient of magnitude 0.6 and 0.8 are shown in MSLP and Z200 regression fields.

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

At low latitudes, the regression of MSLP onto Niño-3.4 reveals the SO (Figs. 3a,d), with enhanced MSLP occurring over cold SST and reduced convection to the north of Australia and reduced MSLP occurring over warm SST and increased convection in the equatorial central Pacific. Comparison with Z200 (Figs. 3b,e) reveals that the tropical component of the SO has, as expected, a baroclinic vertical structure with the tropical Z200 anomalies tending to be of opposite sign to those at the surface. Similarly, the regression onto the DMI reveals the Indian Ocean equivalent of the SO with high (low) pressure occurring in the eastern (western) equatorial Indian Ocean over cold (warm) SST. Examination of the Z200 anomalies reveals that this west–east dipole of surface pressure, associated with the DMI, also has a baroclinic structure in the tropical Indian Ocean.

At extratropical latitudes, the circulation anomalies associated with ENSO and the IOD are equivalent barotropic, taking the form of wave trains emanating from the respective regions of anomalous tropical divergence associated with anomalous tropical convection. For ENSO, the PSA pattern in the eastern Pacific, emanating from upper-level divergence in the equatorial central Pacific, is particularly notable. Although weaker, a Rossby wave train appears to emanate from the suppressed convection in the eastern Indian Ocean/Maritime Continent region, resulting in an equivalent barotropic ridge (anticyclone) over southern Australia. When added to the MSLP anomaly associated with the equatorially trapped baroclinic component, Australia experiences a latitudinally broad region of high MSLP during ENSO. Examination of the regression of MSLP and Z200 onto the DMI reveals that high pressure over southern Australia is primarily associated with the equivalent barotropic wave train driven by convective variations in the eastern Indian Ocean during the IOD. Saji and Yamagata (2003) suggest that such a wave train mechanism is a process whereby the IOD generates temperature anomalies remote to the IOD poles. We will refer to this as the East Indian Ocean (EIO) wave train.

The primary role of convective variations in the IODE region for generating the EIO wave train is confirmed by careful examination of the partial regression onto Niño-3.4 when the effects of the DMI are removed: the ridge over southern portions of Australia during El Niño is significantly weakened (Fig. 3d). Thus, surface high pressure at more southerly latitudes over Australia during El Niño originates primarily from the equivalent barotropic wave train driven by the convective/SST variations to the north and west of Australia rather than from thermally direct, deep baroclinic SO that is the response to rainfall variations in the central Pacific.

The regressions of MSLP and Z200 onto the DMI and Niño-3.4 provide insight into the rainfall anomalies over Australia associated with ENSO and the IOD (Figs. 3c,f,i,l). The primary rainfall response in austral winter during ENSO is confined to the central portions of eastern Australia, which directly feel the anomalous vertical motion of the baroclinic, equatorially trapped SO. The primary response during the IOD is across southern portions of the country, associated with southward displacement of the midlatitude westerlies due to the anomalous equivalent-barotropic ridging over southern Australia. Orographic rainfall effects in eastern Australia seem particularly evident during the IOD in winter, with dry anomalies developing on the western (up wind) side of the Dividing Range where easterly anomalies during the positive phase of the IOD act to reduce the normal westerly flow. Such an impact does not extend all the way to the east coast because easterly anomalies there are conducive to rain-bearing systems impinging from the east (e.g., Hendon et al. 2007); however, these easterly anomalies do not appear to be sufficiently intense for a significant rainfall increase. The partial regression confirms that the rainfall anomalies across southern Australia are more prominent during the IOD when ENSO effects are removed and the southern anomalies during ENSO are weakened when the IOD effects are removed.

This more primary impact of the IOD on rainfall across southern Australia is corroborated by the regression of var500 onto Niño-3.4 and DMI during winter (Fig. 4). Although synoptic storminess is reduced across southern Australia both during El Niño and the IOD, the reduction is more strongly associated with the IOD. And, the modest reduction in storminess across southern Australia during El Niño is largely reduced when IOD effects are removed. This reduction in storminess across southern Australia associated with the IOD is consistent with the modeling study by Ummenhofer et al. (2009) who showed that an equivalent barotropic ridge in response to cold SST forcing in the tropical eastern Indian Ocean acted to reduce the mean-state westerly vertical shear across southern Australia, hence reducing baroclinicity. West–east steering of upstream storms into the region is also reduced. More modest reductions of storminess associated with ENSO are seen in northeast Australia, which we attribute to the enhanced surface pressure/downward vertical velocity variations driven by the SO during El Niño. Also evident in Fig. 4 are the enormous changes in storminess in the central South Pacific associated with the development of the PSA pattern during ENSO.

Fig. 4.
Fig. 4.

Regression of var500 (m2) onto (a) Niño-3.4, (b) partial Niño-3.4, (c) partial DMI, and (d) DMI for austral winter (JJA). The plotting convention is as in Fig. 2.

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

We also note that earlier work by Karoly (1989) shows a significant positive height and MSLP anomaly south of Australia during a wintertime composite of El Niño events, which is absent in our regressions onto the Niño-3.4 index for JJA (Fig. 3). The composite in Karoly (1989) was constructed using four El Niño events starting in 1972, 1976, 1977, and 1982, which were all positive IOD years. Hence, the strong high MSLP anomaly center south of Australia that the Karoly (1989) study ascribed to ENSO is interpreted here to be a response to convection anomalies over the IODE region stemming from the covariation of the IOD. We note that for the period considered in the present study (1979–2008), the IOD and ENSO are not strongly related in winter; hence, we do not associate the surface high to the south of Australia with ENSO.

In addition to the wave trains emanating from the eastern Indian Ocean and central Pacific, the regression and partial regression of MSLP and Z200 onto Niño-3.4 (Figs. 3a,b,d,e) also emphasize an apparent equivalent barotropic wave train emanating from equatorial Africa (EqA), resulting in a strong positive MSLP anomaly center over the southern central Indian Ocean (40°S, 80°E). It appears to be associated with a concentrated OLR anomaly in the EqA and east Africa region (Figs. 2b,h). The impact of this pattern on southern East Africa appears minor, as there is only a weak ENSO-related anomaly over the region (Nicholson and Kim 1997); the impact on southern Australia is also small, as the EqA wave train does not produce an anomaly center over Australia.

Finally, impacts from the IOD and ENSO appear to offset in the IODW region. As part of the zonal hemispheric response of the SO, El Niño tends to induce a positive MSLP anomaly in the IODW region, which is unfavorable for convection and rainfall. But, the development of a positive SST anomaly in the IODW region associated with a positive IOD is conducive to a lower MSLP and enhanced convection (Figs. 2 and 3). As a result, convection over the IODW region is neither correlated with Niño-3.4 nor with IODW SST (Fig. 1b). The absence of an extratropical response emanating from the IODW region appears to be a consequence of this “SO–IODW SST” offsetting effect.

4. Austral spring

Regression anomaly patterns for SON are depicted in Figs. 5, 6, and 7. As we will see below, it is difficult to isolate the effects of ENSO and the IOD, or the impact of IODE from that of IODW, during austral spring because the correlations of Niño-3.4 with the DMI and IODW with IODE are far higher than in JJA (Fig. 1c). And, convection in both the IODW and IODE regions is significantly correlated with Niño-3.4 (Fig. 1d).

Fig. 5.
Fig. 5.

As in Fig. 2 but for austral spring (SON).

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

Fig. 6.
Fig. 6.

As in Fig. 3 but for austral spring (SON).

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

Fig. 7.
Fig. 7.

As in Fig. 4 but for austral spring (SON).

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

The SST and convection/vertical velocity patterns during SON (Fig. 5) share many features with those in JJA, noting now the much stronger similarity between the patterns of anomalies for the IOD and ENSO as compared to JJA (Fig. 2). Partial regression does now, however, show a reduced connection of the IOD to SST/convection anomalies in the Pacific and a weakened connection of ENSO to the SST/convection anomalies in the Indian Ocean. In comparison to JJA, the SST/convection anomalies for both Niño-3.4 and the DMI are notably stronger in SON, reflecting that both ENSO and the IOD tend to be stronger in austral spring than in winter. In the case of Niño-3.4, the convection/vertical velocity anomalies are broader. For instance, the sinking motion over tropical northeast Australia during El Niño, which we ascribe to the direct impact of the SO, now extends southward along the east coast.

The baroclinic tropical responses and the equivalent barotropic extratopical responses are again apparent in MSLP and Z200 (Fig. 6). The baroclinic, equatorially trapped SO and its equivalent in the Indian Ocean are again clearly associated with the respective SST/convection anomalies in each basin. The equivalent barotropic PNA and PSA patterns, emanating from the central Pacific convective anomalies that are predominantly associated with Niño-3.4, are clearly evident. The EIO wave train, emanating from reduced convection in the eastern Indian Ocean, is seen to be more prominently associated with the IOD and, in comparison with winter, its downstream centers southeast of Australia are now stronger. However, the resulting ridge across southern Australia is shifted west compared to winter, perhaps stemming from enhanced convection in the IODW region now acting as a Rossby wave source of the opposite sign to that in the IODE region. Although the SO–IODW SST offsetting effect over the IODW region continues, the influence of the IODW SST appears to now win out so that a wave train emanating from the IODW region is now apparent. A central feature during austral spring is that convection over the IODW region varies coherently with the DMI (Fig. 1d, with a correlation of −0.75) and is, in turn, highly coherent with convection and SST anomalies over the IODE region (Fig. 6). The distinctive extratropical response to the anomalous vertical motions over the IODW region is referred to as the West Indian Ocean (WIO) wave train. It appears to contribute to the anomalous high MSLP and Z200 centers south of Australia that are associated with the EIO wave train (e.g., Figs. 3j,k), but appears to act to shift the high westward compared to winter when only the EIO is apparent. The SO influence on convection in the EqA region is still visible in the Niño-3.4|DMI case (Fig. 6e): an imprint of the EqA wave train is seen (with a high center at 45°S, 80°E) similar to that shown in Fig. 3e, although not statistically significant in the spring season.

With regard to the impact on Australian rainfall, the direct tropical impact of ENSO during austral spring is still confined mainly to the northeastern portions of the continent, while the impact of the IOD is mainly across the higher latitudes. The IOD impact is conducted through the extratropical teleconnection response (the EIO and WIO wave trains), as there is no evidence of a continuous convection anomaly extending from the tropical IODE region into extratropical southern Australia, unlike for JJA (Figs. 2 and 5). In contrast to winter, orographic effects of the IOD on eastern rainfall are less distinct in spring, reflecting the westward displacement of the ridge across southern Australia in spring as compared to winter. The impact of the IOD and ENSO on Australian rainfall are also understood from consideration of impacts on var500 (Fig. 7). As in winter, the ridge associated with the EIO wave train across southern Australia acts to reduce synoptic storminess across southern Australia.

The association of the EIO wave train with ENSO in spring, which accounts for the ENSO–rainfall variations across southern Australia, appears to stem from the strong covariation of the IOD with Niño-3.4. This covariation of the IOD with ENSO results in a strong modulation of convection in the IODE during ENSO. The primary role of convection variations in the IODE region for generating the EIO wave train pattern and associated rainfall variations across southern Australia during ENSO is revealed in Fig. 8, which shows the regression of Australian rainfall onto Niño-3.4 after the effect of OLR in the IODE region is removed. The reduction in rainfall across southern Australia in comparison to the full regression onto Niño-3.4 (Fig. 6c) is marked and even more pronounced than when the effect of the DMI is removed from ENSO (Fig. 6f). This result emphasizes the primary importance of convection variations in the IODE region for driving the EIO wave train during both ENSO and the IOD. It is unambiguous that the impact on southern Australia rainfall is far greater when both ENSO and the IOD effects are considered jointly because they both act to modulate convection in the Indian Ocean region, which is the source of the extratropical response across southern Australia.

Fig. 8.
Fig. 8.

Partial regression of Australian rainfall (mm) onto Niño-3.4 after the effects of OLR in the IODE region have been removed for the austral spring season (SON). The plotting convention is as in Fig. 2.

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

Our central result that the impact of ENSO on southern Australia is conducted through a modulation of convection in the IODE is “confirmed” during recent events in 2007 and 2009. In both events, a positive IOD occurred in conjunction with a La Niña, with southern Australia experiencing anomalously dry conditions (Cai et al. 2009a,c). A typical La Niña event gives an increase in rainfall in the southern Australian region; however, if a positive IOD is present then the accompanying drying effect prevails. Our result suggests the reduced convection anomaly in the IODE region, associated with positive IOD events, prohibited the impact of these two La Niña events from penetrating into southern Australia through the EIO wave train.

5. Discussion and conclusions

Tropical SST variations associated with ENSO and the IOD drive both tropical and extratropical responses. The tropical response is understood as the baroclinic response to the diabatic (convective) heating anomalies induced by the tropical SST anomalies (e.g., Gill 1980). Remote impacts of this response are confined to near-tropical latitudes because this direct baroclinic response is equatorially trapped. The same diabatic heating anomalies that drive the equatorially trapped response also act to excite equivalent barotropic Rossby wave trains that propagate from the tropics into the extratropics (e.g., Hoskins and Karoly 1981). The geographical location of Australia means that it feels effects from both the tropical baroclinic and extratropical barotropic responses during ENSO and the IOD. We have argued that the mechanism by which each type of response affects local climate is different. The direct tropical response during ENSO is the SO, whose impact on Australian rainfall is confined mainly to near-tropical portions of eastern Australia. Rainfall is suppressed during El Niño because near-tropical eastern Australia directly experiences sinking motion and higher surface pressure associated with the western pole of the SO. Although the best known extratropical teleconnections associated with ENSO originate from the convective variations in the central Pacific, higher latitude impacts in Australia during ENSO and the IOD are argued to stem from the equivalent barotropic Rossby wave trains emanating from the tropical Indian Ocean. These wave trains influence southern Australia climate by altering the mean state westerlies, thereby altering the mean state baroclinicity, the west–east steering and orographic rainfall effects.

A primary result of the present study is the identification of the Rossby wave trains and their source regions that develop individually and mutually during ENSO and the IOD. These wave trains and their sources are illustrated schematically in Fig. 9, which is an updated version of Fig. 11 in Karoly (1989). In austral winter, when ENSO and the IOD are not strongly related, the primary wave trains in each case are distinctive and independent. For ENSO the primary wave train is the PSA pattern, which radiates poleward and eastward from the convective anomalies in the central equatorial Pacific and so does not impact Australia. Hence, the impact of ENSO on Australian rainfall in winter stems mainly from the equatorially trapped, baroclinic SO and so is confined primarily to near-tropical latitudes of eastern Australia. The IOD, however, is associated with development of the EIO wave train that emanates from convective variations in the equatorial eastern Indian Ocean and results in an equivalent barotropic ridge over southern Australia during the positive phase of the IOD. The associated easterly anomalies there weaken the rain-bearing eastward-moving weather systems leading to a rainfall reduction over much of southern Australia. Strong orographic effects along the Great Dividing Range of southeastern Australia are also evident with rainfall reductions on the west side during positive phases of the IOD. Interestingly, little rainfall anomaly on the east side of the Dividing Range is seen because easterly anomalies there promote onshore moisture advection from the Tasman Sea, thereby compensating for the rainfall reduction due to reduced storminess.

Fig. 9.
Fig. 9.

Schematic illustration of the typical wave trains associated with the (a) IOD and (b) ENSO for JJA and (c) for the IOD and ENSO together for SON. This description is for positive phases of the IOD and El Niño. Typically opposite patterns occur during negative IOD and La Niña. Shaded blue (red) areas indicate regions of increased (decreased) tropical convection. Blue (red) contours indicate anomalously low (high) upper-level heights. The dashed lines trace the prominent wave trains mentioned in the text: gray for the EIO wave train, green for the WIO wave train, and orange for the EqA wave train, PNA and PSA wave trains not marked but visible in (b) and (c).

Citation: Journal of Climate 24, 15; 10.1175/2011JCLI4129.1

There is little evidence of a wave train emanating from the western pole of the IOD in winter, which we ascribe to the incoherent SST variations in the western and eastern IOD poles at this time of year, together with the impact of the SO. On the other hand, subsidence associated with the western pole of the SO acts to reduce convection over EqA, thus apparently acting as a Rossby wave source for a wave train that produces a strong anomaly center in the midlatitude Indian Ocean. Although this EqA wave train does not directly impact Australia, it does appear to contribute significantly to climate variability of the extratropical Indian Ocean.

In the SON season, ENSO and the IOD are positively correlated; hence, it is difficult to separate their individual effects on the generation of Rossby wave trains. We thus show only one schematic of the Rossby wave trains for ENSO/IOD during austral spring (similar to Fig. 11 of Karoly 1989). The primary wave trains radiating from the convective variations in the central Pacific are the PSA (radiating into the Southern Hemisphere) and PNA (radiating into the Northern Hemisphere). The EIO pattern is still evident, emanating from the IOD east region. However, because SST and convection in the IODW region now strongly covary with the IODE region (but with opposite sign), the IODW region now appears to act as the Rossby wave source for the WIO wave train. The EIO and WIO wave trains appear to share a common anomaly center south of Australia, together leading to a reduction in rainfall across southern portions of Australia during positive IOD/El Niño. In comparison to austral winter, the anomalous high pressure center that develops to the south of Australia during El Niño/positive IOD is shifted westward, perhaps due to the contribution of the wave train of the opposite sign emanating from the IODW region. This has the apparent effect of lessening the orographic impacts of the EIO wave train on rainfall in eastern Australia.

Although our synthesis of the Rossby wave trains and their sources associated with ENSO and the IOD that is provided in Fig. 9 is consistent with previous diagnostic and modeling studies (e.g., Jin and Hoskins 1995; DeWeaver and Nigam 2004), our present results need further confirmation with additional modeling studies. For instance, many previous studies have showed that diabatic heating in the vicinity of the Maritime Continent can excite a Rossby wave train that propagates into the Southern Hemisphere; this is reminiscent of the EIO wave train, although these studies have generally focused on the boreal winter (December–February) and austral autumn seasons (March–May; e.g., Watterson 2010). The sensitivity of the Rossby wave response to heating locations in tropical African and Indian Ocean sectors needs to be explored for austral winter and spring basic states. Also, the notion of additive forcing by the opposite poles of the IOD during austral spring for the height response across southern Australia needs to be verified with imposed dipole heating anomalies. Nonetheless, the present results should form the basis for such additional investigations.

Our results do provide an enhanced understanding of both the mechanisms of the global teleconnections of ENSO and the IOD and the distinctive mechanism by which ENSO and the IOD affect Australian rainfall. These results have other important implications. First, extended range prediction of Australian climate depends not just on the ability to predict the occurrence and phases of ENSO, but also on the ability to predict associated SST/convective variations in the eastern Indian Ocean, which are the primary source of the Rossby wave trains that cause rainfall variations across the southern portions of the country even during ENSO. Hence, extended range prediction of the impacts of ENSO in southern portions of Australia will be limited by the ability to predict the convective variations in the eastern Indian Ocean that act as the Rossby wave source for the remote response in southern Australia. This region of the tropical Indian Ocean is notoriously difficult to predict due to both problems in simulating the mean state (e.g., Fischer et al. 2005) and generally lower intrinsic predictability of surface climate in the Indian Ocean than in the Pacific (e.g., Zhao and Hendon 2009). To advance predictive capability of extratropical Australian climate, even in forecast systems that already skillfully predict ENSO, thus requires improved simulation and prediction of the climate to the northwest of Australia.

Second, impacts of long-term trends in tropical SST on Australian climate will depend on both mean warming and the frequency and magnitude of ENSO and IOD events. For example, any increased frequency of El Niño would be expected to reduce rainfall across central eastern Australia, while impacts across southern Australia would require a trend toward more occurrences of positive IODs. Cai et al. (2009b) have indicated that the frequency of positive IOD events has increased in recent years, as evidenced by a positive trend in the DMI. This observed trend in the DMI is mainly confined to late winter and spring, at which time it accounts for a significant portion of the observed rainfall decline across southern Australia, and has contributed to significant bushfires over southeastern Australia (Cai et al. 2009a). The strongest recent rainfall decline across southern Australia, however, occurred in austral autumn (e.g., Cai and Cowan 2008; Murphy and Timbal 2008), during which the DMI has shown little trend. Thus, the recent rainfall decline across southern Australia during austral autumn is unlikely to have resulted from a trend in IOD activity. Nonetheless, future impacts of trends in tropical SST on Australia climate will depend on the mean state and frequency of the tropical drivers. Projections of the climate of the twenty-first century from many models suggest a trend toward more positive IOD events (Cai et al. 2011), thus indicating the likelihood of future reduction in rainfall across southern Australia. The robustness of these projected trends in the IOD needs to be ascertained.

Acknowledgments

This research is supported by the Australian Climate Change Science Program (ACCSP), Southeast Australia Climate Initiative (SEACI), the Western Australian Marine Science Institution (WAMSI), and the Grains Research and Development Corporation (GRDC). The authors wish to thank Ian Watterson and two anonymous reviewers who provided useful feedback that helped improve this paper. We also thank Joanne Richmond for her assistance in producing Fig. 9. This paper follows from presentations at the Southeast Australia Rainfall Workshop, September 2009, sponsored by the Climate Change Research Centre, UNSW.

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.

    • Search Google Scholar
    • Export Citation
  • Cai, W., and T. Cowan, 2008: Dynamics of late autumn rainfall reduction over southeastern Australia. Geophys. Res. Lett., 35, L09708, doi:10.1029/2008GL033727.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Cai, W., A. Pan, D. Roemmich, T. Cowan, and X. Guo, 2009c: Argo profiles a rare occurrence of three consecutive positive Indian Ocean Dipole events, 2006–2008. Geophys. Res. Lett., 36, L08701, doi:10.1029/2008GL037038.

    • Search Google Scholar
    • Export Citation
  • Cai, W., A. Sullivan, T. Cowan, J. Ribbe, and G. Shi, 2011: Simulation of the Indian Ocean Dipole: A relevant criterion for selecting models for climate projections. Geophys. Res. Lett., 38, L03704, doi:10.1029/2010GL046242.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and S. Nigam, 2004: On the forcing of ENSO teleconnections by anomalous heating and cooling. J. Climate, 17, 32253235.

  • Fischer, A. P., P. Terray, E. Guilyardi, S. Gualdi, and P. Delecluse, 2005: Two independent triggers for the Indian Ocean dipole/zonal mode in a coupled GCM. J. Climate, 18, 34283449.

    • Search Google Scholar
    • Export Citation
  • Frederiksen, C. S., and J. S. Frederiksen, 1996: A theoretical model of Australian northwest cloudband disturbances and Southern Hemisphere storm tracks: The role of SST anomalies. J. Atmos. Sci., 53, 14101432.

    • Search Google Scholar
    • Export Citation
  • Ghil, M., and K. C. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part II: Southern Hemisphere. J. Atmos. Sci., 48, 780790.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447462.

  • Hendon, H. H., D. W. J. Thompson, and M. C. Wheeler, 2007: Australian rainfall and surface temperature variations associated with the Southern Hemisphere Annular Mode. J. Climate, 20, 24522467.

    • Search Google Scholar
    • Export Citation
  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109, 813829.

    • Search Google Scholar
    • Export Citation
  • 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, 11791196.

    • Search Google Scholar
    • Export Citation
  • Jin, F., and B. J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52, 307319.

  • Jones, D., and G. Weymouth, 1997: An Australian monthly rainfall dataset. Tech. Rep. 70, Bureau of Meteorology, 19 pp.

  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events. J. Climate, 2, 12391252.

    • Search Google Scholar
    • Export Citation
  • Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 12751277.

    • Search Google Scholar
    • Export Citation
  • 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, 28722880.

    • Search Google Scholar
    • Export Citation
  • Murphy, B., and B. Timbal, 2008: A review of recent climate variability and climate change in south-eastern Australia. Int. J. Climatol., 28, 859879.

    • Search Google Scholar
    • Export Citation
  • Nicholls, N., 1985: Towards the prediction of major Australian droughts. Aust. Meteor. Mag., 33, 161166.

  • Nicholls, N., 1989: Sea surface temperatures and Australian winter rainfall. J. Climate, 2, 965973.

  • Nicholls, N., 2009: Local and remote causes of the southern Australian autumn–winter rainfall decline, 1958–2007. Climate Dyn., 34, 835845, doi:10.1007/s00382-009-0527-6.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and J. Kim, 1997: The relationship of the El Niño–Southern Oscillation to African rainfall. Int. J. Climatol., 17, 117135.

    • Search Google Scholar
    • Export Citation
  • 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.

    • Search Google Scholar
    • Export Citation
  • Risbey, J. S., M. J. Pook, P. C. McIntosh, C. C. Ummenhofer, and G. Meyers, 2009: Characteristics and variability of synoptic features associated with cool season rainfall in southeastern Australia. Int. J. Climatol., 29, 15951613.

    • Search Google Scholar
    • Export Citation
  • Saji, N. H., and T. Yamagata, 2003: Possible impacts of Indian Ocean Dipole mode events on global climate. Climate Res., 25, 151169.

  • Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360363.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45, 12281251.

    • Search Google Scholar
    • Export Citation
  • Tapp, R. G., and S. L. Barrell, 1984: The north–west Australian cloud band: Climatology, characteristics and factors associated with development. Int. J. Climatol., 4, 411424.

    • Search Google Scholar
    • Export Citation
  • Ummenhofer, C. C., A. S. Gupta, M. J. Pook, and M. H. England, 2008: Anomalous rainfall over southwest Western Australia forced by Indian Ocean sea surface temperatures. J. Climate, 21, 51135134.

    • Search Google Scholar
    • Export Citation
  • Ummenhofer, C. C., A. S. T. A. Sen Gupta, and M. H. England, 2009: Modulation of Australian precipitation by meridional gradients in east Indian Ocean sea surface temperature. J. Climate, 22, 55975610.

    • Search Google Scholar
    • Export Citation
  • Walker, G. T., 1923: Correlations in seasona variations in weather, VIII. Mem. India Meteor. Dept., 24, 75131.

  • Watterson, I. G., 1987: The effect of the Hadley Circulation on the meridional propagation of stationary waves. Quart. J. Roy. Meteor. Soc., 113, 779813.

    • Search Google Scholar
    • Export Citation
  • Watterson, I. G., 2010: Relationships between southeastern Australian rainfall and sea surface temperatures examined using a climate model. J. Geophys. Res., 115, D10108, doi:10.1029/2009JD012120.

    • Search Google Scholar
    • Export Citation
  • Webster, P. J., A. M. Moore, J. P. Loschnigg, and R. R. Leben, 1999: Coupled ocean-atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356360.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., and H. H. Hendon, 2009: Representation and prediction of the Indian Ocean Dipole in the POAMA seasonal forecast model. Quart. J. Roy. Meteor. Soc., 135, 337352.

    • Search Google Scholar
    • Export Citation
Save