To examine the root cause of the convective activity over the Tasman Sea and its role in developing the SST and atmospheric W4 patterns, we conducted sensitivity experiments with the CGCM. The spatial patterns of the second EOF mode for all the simulations show the SST-W4 pattern in the second mode. In the noTIOTP, the SST over the Indo-Pacific region (25°S-25°N) is nudged to the climatological SST to suppress the influence of tropical region (Liess et al. 2014) on Tasman Sea climate variability, so also on the generation of W4 patterns. Interestingly, the EOF pattern remains unchanged, suggesting they are not dependent on the variations of SST in tropical regions. Further, a composite analysis is carried out for CTL and noTIOTP simulation results to investigate the dynamical mechanism. During positive years in the CTL simulation, wave activity flux emanating from the Tasman Sea in response to precipitation and divergent wind contributes to the development of the W4 patterns in the 250 hPa geopotential height, 850 hPa wind, and SST anomalies. In the noTIOTP experiment, the dynamics seen to be similar with the CTL simulation and hence not affected by the tropical climate. In fact, the subtropical dipoles in each basin can be generated owing to the SAM in the absence of tropical climate variability (Morioka et al. 2014). Further, the geopotential height does not vary over mid and high latitudes in the noTIOTP experiment, omitting the role of the SAM in SST and atmospheric W4 patterns. Also, the convective activity over the Tasman Sea is unaffected from the tropical region. The scenario is the opposite during negative years. Therefore, it can be hypothesized that the variability over the Tasman Sea might be locally generated. To confirm this, we perform the noTS experiment by nudging the SST to its climatological value over the Tasman Sea (black rectangle box). The decoupling of air-sea interaction over the Tasman Sea supresses the SST and atmospheric W4 patterns in the southern subtropics, especially over the Indo-Atlantic Ocean. The SST anomalies over the eastern Pacific and Atlantic Oceans are not sensitive to the Tasman SST variability, as seen in the Figure.  However, in the absence of convection activity over the Tasman Sea,  the circumglobal SST and atmospheric W4 patterns disappear. Hence the noTS experiment concludes that variability in the Tasman Sea as a necessary condition for forming the Rossby wave and, subsequently, the SST and atmospheric W4 patterns. 

At 24-month lag (Figure a), the SST pattern shows the positive phase of South Pacific Meridional Mode (SPMM, i.e., warm SST anomaly in 25°–35°S latitude belt) over the South Pacific Ocean. At the same time, a cold (warm) SST anomaly is noticed south of South Africa, southeast Indian Ocean, and southwest Atlantic Ocean. The anomalous geopotential height and related wind strongly follow the SST anomaly over the western subtropical Pacific Ocean extending westward up to the south of Australia as compared to other regions. As a result, a large circulation cell is observed at this lag. At 18-month lag (Figure b), the warm SST anomaly over the South Pacific Ocean weakens leaving footprints over three regions: (a) South of Australia, (b) the central South Pacific Ocean, and (c) the eastern South Pacific Ocean. The warm SST anomaly south of Australia is mainly due to the meridional wind anomaly (Figures a and b) and sustains through coupling with the atmospheric circulation over Australia. Gradually, the geopotential height strengthens over the southern central Pacific Ocean, eastern South Pacific Ocean, southwest Atlantic Ocean, the southern part of the South Africa continent, and the southeast Indian Ocean. After 6 months (at T = −12, Figure c), the equatorward movement of cold SST anomaly along with the negative geopotential height anomaly and associated wind is noticeable southeast of New Zealand. Similarly, the west coast of South America has a cold SST anomaly. The negative anomalies deepen and expand equatorward in subsequent months (Figures d and e) and evolves as an SST W4 pattern in the southern mid-latitudes. In this way, the SST W4 pattern evolves with the help of SST residuals from the SPMM by the footprinting mechanism in 2 years. The background SST residuals favor a more frequent positive/negative SST W4 pattern which leads to its decadal variability with positive/negative phase. Watch this video for illustration of this mechanism: