Climate prediction for Brazil's Nordeste by empirical and numerical modeling methods

Abstract

AbstractClimate prediction for Brazil's Nordeste meaningfully uses information through January to forecast the rainfall of March–June. Empirical methods developed at the University of Wisconsin use as predictors preseason (October–January) rainfall in the region and January indices of the fields of meridional wind component and sea surface temperature (SST) in the tropical Atlantic and the equatorial Pacific, as input to stepwise multiple regression to predict the March–June rainfall at a network of 27 stations. The training period is 1921–1957, the verification period 1958–1989, and real‐time forecasting continued to 2000. A numerical model (ECHAM4.5) experiment conducted at International Research Institute for Climate and Society at Columbia University used global SST information through January to predict the March–June rainfall at three gridpoints in the Nordeste for the years 1968–1999. For the same predictand, a complementary empirical experiment was conducted using the gridded rainfall observations, again with training period 1921–1957, to yield predictions for 1968–1999. Over this period predicted versus observed rainfall are evaluated in terms of correlation, root‐mean‐square error, absolute error, and bias. Compared with the empirical method the numerical modelling produces larger errors and negative bias. The empirical method captures 59% and the numerical modelling 49% of the variance. Copyright © 2008 Royal Meteorological Society