Search

The article reports on the study that determined the climate prediction for Brazil's Nordeste to forecast the March to June rainfall. The researchers conducted a complementary empirical experiment using the gridded rainfall observations, with training period of 1921-1957, to yield predictions for 1968-1999. It also evaluates the predicted versus observed rainfall over the period in terms of correlation, absolute error, mean-square error and bias. It is indicated that the empirical method captured 59%, while the numerical modelling obtained 49% of the variance.