Impactos dos eventos La Niña plurianuais no clima da América Do Sul

Abstract

The effect of multiyear La Niña (LN) events on the climate of South America (SA), mainly on precipitation and air temperature, was assessed considering composite analyses of the ten persistent LN events over two successive years, referred to as Y1 and Y2 for the 1901–2012 period. The results revealed that the teleconnections related to a multiyear LN event during its Y1 and Y2 years were associated with different tropical ocean conditions. In spring, LN conditions related to the negative Sea Surface Temperature (SST) dipole in the Indian Ocean during the LN (Y2) were not found for LN (Y1). The LN conditions were associated with different SST patterns in the tropical Atlantic for the other seasons. Over northern/northeastern SA the positive precipitation anomalies became weaker (stronger) during austral summer and autumn (winter and spring) of the LN (Y2) than LN (Y1) and were associated with changes in the Walker cells. Drier conditions were observed during the LN (Y2) than LN (Y1) events over southeastern SA in the austral spring and summer. In the spring of LN (Y2), two Rossby wave trains, one associated with LN-related anomalous cooling in the equatorial Pacific and another triggered by the upper-level anticyclone in the tropical Indian Ocean, characterized the circulation pattern over SA that explains the difference in precipitation anomalies between the LN (Y2) and LN (Y1). Regarding the air temperature variations for both LN, the negative anomalies were noted in the northwest region, associated with the cold advection from the Equatorial Pacific in response to LN and the increase in precipitation during the winter and spring austral. In turn, a cooling (warming) pattern was observed in the northwest (central east) of the tropical AS during the summer and over the entire territory in the autumn for the LN (Y1), associated with the cold advection over the tropical AS and increased precipitation over these regions. Similar patterns are found for LN (Y2) but this pattern was weakened. These results indicate that accurate prediction of LN duration is crucial in a climate monitoring context.