Use of ERA5 to Initialize Ensemble Re-forecasts

This study summarizes the results from re-forecast experiments initialized with ERA5. Results suggest that initializing the re-forecasts from ERA5 instead of ERA Interim improves the extended-range skill scores up to week 4. The use of the ERA5 EDA to perturb the re-forecast initial conditions has been assessed. The ERA5 EDA provides flow-dependent EDA initial perturbations across the re-forecast years instead of the non-flow-dependent perturbations in the current operational set-up.  Results show that it is beneficial to use the ERA5 EDA, even if the impact is limited to just the first week of the re-forecasts. Different strategies for land and wave re-forecast initialization have been tested. The consistency of the re-reforecast land surface climatology with real-time forecasts is improved when initializing from ERA5 and, as a consequence, no separate offline land simulation would be needed to initialize the land in the ERA5 reforecasts.