The South American rainfall Dipole (SAD) is a renowned spatial structure present in the austral summer as part of the South American monsoon system. SAD phases have been related with extreme precipitations and severe droughts across South America, but are yet to be predicted. Here, we reveal $2$ robust and reliable intraseasonal windows in the accumulated SAD index where we can forecast its quantile-state between $5$ to $15$ and $60$ to $70$ days in advance ($99\%$ significance level). These windows are insensitive to variations in the pole’s size and accumulation window, and results are consistent across different quantiles states (median, tercile, and quartile). Our method, which is based on analysing the lagged mutual information between future and present states, could be used in the development of early-warnings for extreme rainfall events. Moreover, it is unrestricted to the present analysis, being applicable to other stationary signals where a forecast is missing.

This study aims at understanding the impact of low-frequency climate modes on Rossby Wave Packets (RWPs) during the southern hemisphere summer. In particular, we focus on long-lived RWPs (lifespan above 8 days) and determine how El Niño-Southern Oscillation (ENSO) and Southern Annular Mode (SAM) modulate their frequency of occurrence plus the main areas of detection and dissipation. We find that the occurrence of long lived RWPs is maximum during El Niño years and negative SAM events. Years with largest numbers of long-lived RWPs are characterized by a zonally symmetric and narrow upper level jet that is shifted northward from its climatological position. Conversely, when the jet is shifted southward, as during positive SAM phases, particularly in the southwestern Pacific basin, the number of long-lived RWPs detected diminishes. El Niño sets atmospheric conditions that support the formation of long lived RWPs whereas La Niña years presents high interannual variability in the frequency of occurrence. Moreover, during El Niño events the main formation area is between 61-120ºE and its main dissipation area between 300-359ºE. During La Niña events, the main formation area moves to 241-300ºE and no main dissipation area is identified. During positive SAM two main formation areas appear at 61-120ºE and 241-300ºE and two main dissipation areas between 121-180º and 301-359º, whereas in negative SAM only one formation area at 241-300º is detected and no main dissipation area is found.

The Madden-Julian Oscillation (MJO) is a major source of predictability on the sub-seasonal (10- to 90-days) time scale. An improved forecast of the MJO, may have important socioeconomic impacts due to the influence of MJO on both, tropical and extratropical weather extremes. Although in the last decades state-of-the-art climate models have proved their capability for forecasting the MJO exceeding the 5 weeks prediction skill, there is still room for improving the prediction. In this study we use Multiple Linear Regression and an Artificial Neural Network as post-processing methods to improve one of the currently best dynamical models developed by the European Centre for Medium-Range Weather Forecast (ECMWF). We show that the post-processing with the machine learning algorithm employed leads to an improvement of the MJO prediction. The largest improvement is in the prediction of the MJO geographical location and intensity.