Interannual Variability and Seasonal Evolution of Summer Monsoon Rainfall in South America

Abstract

The analysis of the interannual variability of the South American monsoon rainfall is carried out separately for austral spring and summer (and for November and January), based on a 40-yr station gauge dataset. Relationships between modes of variability in these seasons show the influence of antecedent conditions in spring (or November) on the evolution of the monsoon rainfall in peak summer (or January). In spring the first mode is dipolelike, with opposite loadings over central-eastern and southeastern South America. It is connected with ENSO. The second mode shows the highest loadings slightly south of the South Atlantic convergence zone. The leading mode of summer also features dipolelike oscillations between central-eastern and southeastern South America, but is not strongly connected with ENSO. The second mode represents the impact of ENSO, and the third is modulated by SST anomalies in the southern tropical Atlantic. Significant relationships are disclosed between the first dipolelike modes of spring and summer rainfall and thus between the rainfall in spring and summer over central-eastern South America, which includes part of the monsoon core region. These dipolelike modes are associated with a rotational anomaly over southeast Brazil that either conveys moisture flux into central-eastern Brazil (if it is cyclonic) or into southeastern South America (if it is anticyclonic). In spring this anomaly seems to be remotely forced, but after strong rainfall anomalies over central-eastern Brazil in spring, it tends to reverse sign in peak summer, inverting the dipolelike rainfall anomalies. This reversal is hypothesized to be locally forced by surface?atmosphere feedback triggered by the spring anomalies, as weaker teleconnections in summer allow local processes that are stronger in summer to overcome remote forcing. SST and circulation anomalies associated with the first modes in spring and summer, and also the relationship between the first summer mode and surface temperature in spring, are consistent with that hypothesis.