IMPACTO DA UTILIZAÇÃO DE PREVISÕES "DEFASADAS" NO SISTEMA DE PREVISÃO DE TEMPO POR CONJUNTO DO CPTEC/INPE

THE IMPACT OF USING LAGGED FORECASTS ON THE CPTEC/INPE ENSEMBLE PREDICTION SYSTEM In this work we report the application of the lagged average forecasting technique to CPTEC/INPE ensemble forecast. The CPTEC/INPE data consist of two months samples of 15 days forecast for the variables: geopotential height at 500 hPa, air temperature at 850 hPa and mean sea level atmospheric pressure. We focus on the following: 1) Does the lagged averaged ensemble forecast improve forecast skill compared to the CPTEC/INPE operational ensemble? 2) Is the dispersion of the ensemble useful in predicting forecast skill? The results indicate that the utilization of 12h-lagged average forecasts improves the performance of the operational ensemble contributing to increase the ensemble spreading and, consequently, to reduce the under-dispersion of the system. Also we observed that lagged average forecast (LAF) shows similar performance of the operational EPS-CPTEC/INPE and that there is a tendency to higher performance when spread forecast is low, for 5 and 7 day forecast. These results provide the basis for the operational implementation of the LAF technique, which has low computational cost, and contribute to a more efficient utilization of the CPTEC/INPE ensemble predictions.