How well do we understand the motion of mid-latitude surface cyclones relative to the mean flow ?

Mid-latitude surface cyclones are commonly observed to move across the mean flow from the equator to the pole. Some of them undergo a strong and rapid deepening when they cross the axis of the low-frequency (periods greater than 8 days) jet stream that may lead to damaging winter storms.

After a brief review on the existing literature, the presentation will describe a recent theory that may explain this cross-jet motion which is a generalization of the so-called beta drift in the mid-latitude baroclinic context. Results are supported by idealized numerical experiments performed with the two-layer quasi-geostrophic model. The numerical framework consists in initializing synoptic-scale finite-amplitude cyclones at the lower and upper layer in various baroclinic background flows with increasing spatial complexity. The basic mechanism will be first shown for horizontally uniform zonal flows and then for meridionally confined zonal jets. The key parameter controlling the movement of a surface cyclone across the mean tropospheric zonal jet is shown to be the vertical-average potential vorticity (PV) gradient associated with the background flow. Then, synoptic-scale cyclones are embedded in a background meandering jet where the deformation associated to the jet is shown to modulate the cross-jet motion of the lower-layer cyclones, which is primarily due to the barotropic PV gradient.

The second part of the study consists in validating the theory by analyzing the real case of the European storm Xynthia (26–28 February 2010). To do so, numerical sensitivity experiments using the Météo-France global operational forecast model ARPEGE-IFS were performed. A PV-inversion tool is used to modify the vertical-average PV gradient at the initial time. As expected from the theory, when the PV gradient is intensified, there is a quicker displacement of the surface cyclone toward the jet axis and the jet-crossing phase occurs earlier than in the control forecast. The opposite occurs for a reduced PV gradient.