Application of seasonal climate forecasts in agricultural crop monitoring in Brazil

Abstract:

This work is investigating the contribution of seasonal climate forecasts of Eta regional climate model to support crops in Brazil. The weather conditions are directed related with the crop yield, being a basic parameter for its forecast. The southern region has a subtropical climate and is the major national producer of rice and wheat and also is the second one for soybean, bean and corn. The Eta seasonal forecast model data for southern Brazil was evaluated from 2001 to 2010. Observed data from National and state meteorological agencies were used to evaluate the monthly model performance. The model performance was evaluated by calculating two parameters. The Root Mean Square Error (RMSE) was used to evaluate the monthly forecast averages and the observed precipitation standard deviation. The Skill Score Climatology (SSC) was used to compare the accuracy between the forecast and the climatology. The RMSE showed that in some locations the predicted values by the model were closer to the observed. The SSC showed a systematic error for the predicted values by the Eta seasonal model. This behavior indicates that the climatological analysis is more accurate to predict the monthly climate than the ETA model forecast. Also the consecutive negative bias was observed in some locations that can be corrected removing the systematic error.