Evolution of consistent rainfall anomalies over Brazil during the ENSO cycle

iable. Most of the interannual variability of rainfall in Brazil is associated with the ENSO cycle. However, due to the different rainfall regimes and associated mechanisms, to the evolving sea surface temperature anomalies, and to the changing atmospheric basic state, the timing of the most consistent and strongest anomalies varies throughout the country. Different regions show consistent anomalies in different phases. The mean magnitude, the consistency, and the temporal evolution of rainfall anomalies during the various phases of the ENSO cycle are presented in this study, based on data series from more than 1000 stations all over Brazil. These data series contain at least 5 El Nino and 4 La Nina episodes, during the period 1956-1992.

The analysis is based on composites of seasonal precipitation percentiles, along with the assessment of the consistency of the precipitation anomalies for each individual station, based on the hypergeometric distribution. The consistency analysis for individual stations is much more demanding than that using area averaged precipitation.

Although the ENSO/rainfall relationship in the southernmost and northernmost parts of Brazil have been documented, it is convenient to show the evolution of consistent rainfall anomalies for both, the El Nino and La Nina episodes, all over the country. The present study shows that the impact in South and North Brazil reaches its peak in different seasons, and indicates the sub-regions where it is most consistent. Furthermore, it shows that, although El Nino events had been associated with dryness in Northeast Brazil, this signal is not consistent for the wet season there, but in small areas. On the other hand, the consistent wet anomalies during La Nina episodes are more extended. Other interesting aspects disclosed are the significant anomalies in Central Brazil, in winter (0) (dry season), and the impact in the monsoon season, there and near the South Atlantic convergence zone (SACZ), both during El Nino and La Nina events. The anomalies in the SACZ region change sign from spring to summer.

Although some regions have less data coverage than other ones, the consistent evolution of the anomalies indicates that the picture presented is re