Relationships between the Antarctic Oscillation, the Madden-Julian Oscillation and ENSO, and consequences for rainfall analysis

International audienceThe Antarctic Oscillation (AAO) is the leading mode of atmospheric variability in the Southern Hemisphere mid- and high latitudes (south of 20°S). In this paper, we examine its statistical relationships with the major tropical climate signals at the intraseasonal and interannual timescales, and their consequences on its potential influence on rainfall variability at regional scales. At the intraseasonal timescale, although the AAO shows its most energetic fluctuations in the 30-60-day range, it is not unambiguously related to the global-scale Madden-Julian Oscillation (MJO) activity, with in particular no coherent phase relationship with the MJO index. Moreover, in the high southern latitudes, the MJO-associated anomaly fields do not appear to project coherently on the well-known AAO patterns, and are never of an annular nature. At the interannual timescale, a strong teleconnection with ENSO is found during the peak of the austral summer season, corroborating previous studies. El Niño (La Niña) tends to correspond to a negative (positive) AAO phase. The results are statistically significant only when the seasonal mean fields averaged for the November through February season are considered. Based on these results, we then isolate the specific influence of the AAO on rainfall variability at both intraseasonal and interannual time-scales. The example taken here is Southern Africa, a region under the influence of both the MJO and ENSO, experiencing its main rainy season in austral summer, and containing a relatively dense network of rain-gauge measurements. At the interannual timescale, the significance of the teleconnections between South African rainfall and the AAO reveals itself to be a statistical artifact, and become very weak once the influence of ENSO is removed. At the intraseasonal timescale, the AAO is seen to significantly affect the rainfall amounts over much of the country, without interference with other modes of variability. Its influence in modulating the rain appears to be strongest during La Niña years