On the performance of the Canadian Regional Ensemble Kalman Filter

As a first step towards designing a specific data assimilation system for the upcoming Canadian Regional Deterministic Prediction System (RDPS) at 2.5 km , a Regional Ensemble Kalman Filter (REnKF) is being developed. This system currently runs at 15 km horizontal resolution with 192 members. This REnKF system will provide ensemble members to be used eventually by the Regional Ensemble Variational Analysis (REnVar).

The cycling procedure of the REnKF together with its design are compared with the operational global EnKF system. Results on finite autonomous cycling with REnKF will be presented and inter-compared with the global EnKF. Added values of the REnKF over the global EnKF especially for improvements in Mesoscale circulations, precipitation distribution and tropical storm tracks and intensities will be presented.