Cross-timescale Interference and Rainfall Extreme Events in South Eastern South America

The physical mechanisms and predictability associated with extreme daily rainfall in South East South America (SESA) are investigated for the December-February season. Through a k-mean analysis, a robust set of daily circulation regimes is identified and then it is used to link the frequency of rainfall extreme events with large-scale potential predictors at subseasonal-to-seasonal scales. This basic set of daily circulation regimes is related to the continental and oceanic phases of the South Atlantic Convergence Zone (SACZ) and wave train patterns superimposed on the Southern Hemisphere Polar Jet. Some of these recurrent synoptic circulation types are conducive to extreme rainfall events in the region through synoptic control of different meso-scale physical features and, at the same time, are influenced by climate phenomena that could be used as sources of potential predictability. Extremely high rainfall (as measured by the 95th- and 99th-percentiles) is preferentially associated with two of these weather types, which are characterized by moisture advection intrusions from lower latitudes and the Pacific; another three weather types, characterized by above-normal moisture advection toward lower latitudes or the Andes, are preferentially associated with dry days (days with no rain). The analysis permits the identification of several subseasonal-to-seasonal scale potential predictors that modulate the occurrence of circulation regimes conducive to extreme rainfall events in SESA. It is conjectured that a cross-timescale interference between the different climate drivers improves the predictive skill of extreme precipitation in the region. The potential and real predictive skill of the frequency of extreme rainfall is then evaluated, finding evidence indicating that mechanisms of climate variability at one timescale contribute to the predictability at another scale, i.e., taking into account the interference of different potential sources of predictability at different timescales increases the predictive skill. This fact is in agreement with the Cross-timescale Interference Conjecture proposed in the first part of the thesis. At seasonal scale, a combination of those weather types tends to outperform all the other potential predictors explored, i.e., sea surface temperature patterns, phases of the Madden-Julian Oscillation, and combinations of both. Spatially averaged Kendall's ? improvements of 43% for the potential predictability and 23% for realtime predictions are attained with respect to standard models considering sea-surface temperature fields alone. A new subseasonal-to-seasonal predictive methodology for extreme rainfall events is proposed, based on probability forecasts of seasonal sequences of these weather types. The cross-validated realtime skill of the new probabilistic approach, as measured by the Hit Score and the Heidke Skill Score, is on the order of twice that associated with climatological values. The approach is designed to offer useful subseasonal-to-seasonal climate information to decision-makers interested not only in how many extreme events will happen in the season, but also in how, when and where those events will probably occur. In order to gain further understanding about how the cross-timescale interference occurs, an externally-forced Lorenz model is used to explore the impact of different kind of forcings, at inter-annual and decadal scales, in the establishment of constructive interactions associated with the simulated "extreme events". Using a wavelet analysis, it is shown that this simple model is capable of reproducing the same kind of cross-timescale structures observed in the wavelet power spectrum of the Nino3.4 index only when it is externally forced by both inter-annual and decadal signals: the annual cycle and a decadal forcing associated with the natural solar variability. The nature of this interaction is non-linear, and it impacts both mean and extreme values in the time series. No predictive power was found when using metrics like standard deviation and auto-correlation. Nonetheless, it was proposed that an early warning signal for occurrence of extreme rainfall in SESA may be possible via a continuous monitoring of relative phases between the cross-timescale leading components.