Toward Higher-Resolution Limited-Area Ensemble-Based Data Assimilation for NWP at Environment and Climate Change Canada

ECCC is currently developing a limited-area ensemble system (grid spacing of 10 km) to provide flow-dependent background error estimates at a higher resolution than ECCC’s operational global ensemble Kalman filter (EnKF; grid spacing of 39 km) for deterministic NWP (grid spacing of 2.5 km) based on an Ensemble-Variational (EnVar) data assimilation algorithm. Various ensemble generation approaches were explored and compared so far. A limited-area version of ECCC’s EnKF in combination with various hybrid gain configurations, where a weighted mean of the EnKF ensemble-mean and the EnVar analysis is used to recenter the EnKF analysis ensemble, were first tested. More recently, a local ensemble transform Kalman filter (LETKF) was developed and applied in this limited-area context. Recentering approaches had a significant positive impact on the ensemble forecast scores, but had little effect on the deterministic forecasts when they are used in EnVar. Further impacts of the different approaches on the performances of the deterministic forecasts as well as from an ensemble forecasting perspective will be shown. The impact of having limited-area ensemble perturbations correlated at different degree with the ensemble perturbations at the lateral boundaries will also be presented.