Impact of stochastic cumulus on the NOGAPS ET ensemble forecasting system. Part I: Description and general performance

A simple stochastic cumulus convection scheme is tested within the framework of the new NOGAPS Ensemble Transform (ET) ensemble forecasting system. We will briefly explain the stochastic physics scheme and then show the impact of the addition of the stochastic convection upon ensemble performance under several measures. The addition of the stochastic convection improves performance of the ensembles in the tropics under a variety of metrics, including increased ensemble variance, decreased ensemble mean error, reduction of excessive outliers, and improved Brier scores and ensemble variance-forecast error variance relationships. As the ET is a cycling scheme, the addition of stochastic convection results in both a direct change to the forecasts, and an indirect change to the initial perturbations. Experiments show that improvements are due to both the direct and indirect changes at early forecast times. At later forecast times, most of the improvement can be obtained through the direct changes to the forecast alone.