On the forecast of the onset and end of the convective season in the Amazon

Some features of the climate system that can be considered predictors of the onset and end of the convective season over the Amazon were identified using one-month lag correlations and field composites. The fields analyzed were sea surface temperature (SST), outgoing long-wave radiation (OLR), vertical velocity and upper tropospheric winds. Warm (cold) anomalies in the SST in the tropical North Atlantic and the Caribbean Sea tend to be associated with delayed (early) onsets. Likewise, there is a tendency towards a delayed (early) end of the convective season with cold (warm) anomalies in these ocean regions. In addition, the SST in the cold tongue region of the equatorial Pacific is negatively, though weakly correlated with the onset date. The signal of this SST is more evident in the case of the end date, which is earlier with respect to its mean date in most of El Nino cases. The convective activity intensity itself conditions the onset and the end of the convective season, as it is evidenced by the behavior of the OLR and the vertical velocity fields. The more (less) intense the convective activity over South America during the preceding month, the earlier the onset and the later the end of the convective season on the Amazon region. The prediction of the onset and end dates of the convective season in the Amazon region was explored using a simple multiple regression technique based on the variables that have shown precursor signals with respect to these dates. The correlation coefficient between the predicted and the observed onset date is 0.81, and in the case of the end date, it is 0.76. The skill to predict early, delayed and normal categories was high, since in more than two thirds of the cases the category was successfully predicted, and there were no predictions of categories opposed to those observed.