Abstract: In an ensemble Kalman filter, when the analysis update of an ensemble member is computed using error statistics estimated from an ensemble that includes the background of the member being updated, the spread of the resulting ...

Abstract: The ability of a stochastically perturbed parameterization (SPP) approach to represent uncertainties in the model component of the Canadian Global Ensemble Prediction System was demonstrated in Part I of this investigation. ...

Abstract: Accurately representing model-based sources of uncertainty is essential for the development of reliable ensemble prediction systems for NWP applications. Uncertainties in discretizations, algorithmic approximations, and ...