Understanding the predictability of seasonal precipitation over northeast Brazil

Authors

  • VASUBANDHU MISRA

    1. Center for Ocean–Land–Atmosphere Studies, Institute of Global Environment and Society, Inc., 4041 Powder Mill Road, Suite 302, Calverton, MD 20705, USA
    Search for more papers by this author

Corresponding author.
e-mail: misra@cola.iges.org

ABSTRACT

Using multiple long-term simulations of the Center for Ocean–Land–Atmosphere Studies (COLA) atmospheric general circulation model (AGCM) forced with observed sea surface temperature (SST), it is shown that the model has high skill in simulating the February–March-April (FMA) rainy season over northeast Brazil (Nordeste). Separate sensitivity experiments conducted with the same model that entails suppression of all variability except for the climatological annual cycle in SST over the Pacific and Atlantic Oceans reveal that this skill over Nordeste is sensitive to SST anomalies in the tropical Atlantic Ocean. However, the spatial pattern of SST anomalies in the tropical Atlantic Ocean that correlate with FMA Nordeste rainfall are in fact a manifestation of El Niño Southern Oscillation (ENSO) phenomenon in the Pacific Ocean.

This study also analyzes the failure of the COLA AGCM in capturing the correct FMA precipitation anomalies over Nordeste in several years of the simulation. It is found that this failure occurs when the SST anomalies over the northern tropical Atlantic Ocean are large and not significantly correlated with contemporaneous SST anomalies over the eastern Pacific Ocean. In two of the relatively large ENSO years when the model failed to capture the correct signal of the interannual variability of precipitation over Nordeste, it was found that the meridional gradient of SST anomalies over the tropical Atlantic Ocean was inconsistent with the canonical development of ENSO. The analysis of the probabilistic skill of the model revealed that it has more skill in predicting flood years than drought. Furthermore, the model has no skill in predicting normal seasons. These model features are consistent with the model systematic errors.

Ancillary