A dynamical perspective on predictions of severe European cyclones: the role of large-scale conditions

Severe windstorms regularly inflict damage throughout Europe. This research examines a set of 31 midlatitude cyclones, investigating each in depth but encompassing a wide variety of such storms. The set is selected using the Storm Severity Index, and categorised using two methods. The first method is based on the relationship between the storm and the jet stream. Four groups emerge: storms that move along the edge of the jet; that cross the jet stream early; that cross later; or have a split jet. The second method is based on the Pressure Tendency Equation. This establishes whether horizontal temperature advection or diabatic processes dominate during storm development. These two approaches are linked: storms in the first two jet groups tend to be driven by horizontal temperature advection, and the other two groups by diabatic processes.
This work then studies the storms’ forecast quality and spread using ECMWF data. It finds that storm intensity tends to be under-forecast, the forecast storms move too slowly and are too far south. Forecast quality improves and spread decreases earlier in storms that cross the jet early, compared to those that cross later, suggesting a link between jet interaction and forecast error. Storms where horizontal temperature advection dominates are on average less well forecast than their diabatic counterparts, but diabatically driven storms tend to have greater forecast spread.
Finally, this study proposes metrics for storm-prone situations, examining the configuration of the atmosphere prior to the development of the storms. These describe a variety of key factors for cyclogenesis, such as baroclinicity, barotropicity and moist stability. Of the 31 storms, 29 are associated with a value greater than the 98th percentile of one or more metrics. There is a large overlap between storms where baroclinicity is strong and those where horizontal temperature advection dominates the deepening, confirming that the two approaches are dynamically linked. This relationship between the storms, the dynamics, and the metrics will allow future work to identify sources of uncertainty in modelling severe European windstorms.