Predictability of the stratospheric polar vortex in the ECMWF S2S reforecasts

Extreme stratospheric events, e.g strong vortex events and sudden stratospheric warming (SSW) events, are often the main focus of stratospheric predictability studies. Other than strong vortex and SSW events, strong vortex acceleration and deceleration events are related but less studied events. A better understanding of the mechanisms of acceleration and deceleration events would also contribute to the understanding of SSWs and strong vortex events in the stratosphere. As SSWs tend to be less predictable than strong vortex events, it is hypothesized that the predictability of acceleration and deceleration events might differ as they are related to opposite mechanisms. We identify wind acceleration and deceleration events using the daily mean of the zonal mean zonal winds at 60°N and 10 hPa from the ERA-interim reanalysis for the winters of 1998/99-2018/19. Acceleration and deceleration events are defined as a wind change over a 10-day window above the 60th percentile of the magnitude of all identified events. To evaluate the predictability of the events, the ECMWF S2S hindcasts are verified against ERA-interim data. As expected, the predictability of the events increases with decreasing lead time (as the model initialisation date approaches the event onset date). We also find that all 4 types of events, namely acceleration, deceleration, strong vortex and SSW events, show the same predictability behavior, that is, that the predictability of an event is independent of its nature but dependent only on its magnitude. We discuss the difficulties of the model in predicting events associated with strong wind changes by investigating the heat flux-wind relationship in the model. A better understanding of the predictability and dynamical variability in the stratospheric polar vortex by the model could provide a better understanding of the mechanisms of stratospheric events, thus potentially also improving surface weather predictability.