Tuesday, 25 January 2011: 11:15 AM
613/614 (Washington State Convention Center)
In the current NCEP global ensemble forecast system, the initial perturbations are generated by the Ensemble Transform with Rescaling (ETR) method. We will test the ensemble forecasts using EnKF with all NCEP operational observations. Similar to the operational ETR based system, the ensemble perturbations will be re-centered around the production analysis generated by GSI. Since the EnKF intends to minimize the analysis error variance, it may not be optimal for ensemble forecasts especially at medium-ranges. We will re-organize the initial perturbations in order to optimize the forecasts. This extra step will preserve the initial ensemble covariance structure and will not deminish the performance of the analysis generated by EnKF. Comparisons will be made agaist the operational ensemble forecasts by ETR.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner