Simulated precipitation response to SST forcing and potential predictability in the region of the South Atlantic convergence zone

The sensitivity of the precipitation response in the South Atlantic convergence zone (SACZ) to sea surface temperature (SST) anomaly is investigated by an inter-model comparison study of ensembles of multidecadal integrations of two atmospheric general circulation models (AGCMs)—version 1 of the NASA Seasonal-to-Interannual Prediction Project (NSIPP-1) model, and the NCAR community climate model (CCM3) version 3.6.6. Despite the different physical parameterizations, the two models consistently show an SST-forced signal located mainly over the oceanic portion of the SACZ. The signal has interannual-to-decadal timescales, and consists of a shift and strengthening of the SACZ toward anomalous warm waters. A potential predictability analysis reveals that the maximum predictable variance is about 50% of the total SACZ variance over the ocean, but the signal attenuates rapidly toward the South American continent. This result implies that the land portion of the SACZ is primarily dominated by the internal variability, thereby having a limited potential predictability at seasonal timescales.