Ensemble error growth and the mesoscale predictability of cyclogenesis in the lee of the European Alps

One area of early mesoscale predictability research in a full physics NWP model involved the predictability of cyclogenesis in the lee of the European Alps. Otherwise known as lee-cyclogenesis, it has been suggested that these events experience extended predictability due to there association with topography. Despite these potentially promising results for NWP, recent research has indicated that alpine scale terrain induced flows, such as mountain wave breaking and downslope winds have an inherently shorter prediction period. Furthermore, it has been shown that orographically induced alpine scale phenomenon, such as potential vorticity banners, can contribute to the early stages of lee-cyclogenesis. An ensemble data assimilation and prediction system will be used to explore the impact of initial condition uncertainty on the perturbation error growth in lee cyclones. Emphasis will be placed on the evolution of perturbations during the initial stages of cyclogenesis when the terrain induced features are most relevant, as well as on the upscale growth of these perturbations that influence broader synoptic scales.

In this presentation, an ensemble Kalman filter (EnKF) data assimilation system is used with the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS®) to generate an 0(100) member ensemble of initial condition perturbations for several cyclogenesis events in the lee of the Alps. The perturbation error growth over a 48-hr forecast period is use to approximate forecast uncertainty of lee-cyclogenesis due to initial condition uncertainty. Initial results suggest that the mesoscale location of cyclogenesis and the subsequent track can vary by O(100 km's), with several members generating no cyclone at all. This has significant implications for sensible weather prediction such as precipitation and flooding. Horizontal resolutions ranging from 45- to 5-km are used to explore the perturbation error growth from the synoptic scale down to the Alpine scale. The role of topography on the interaction of scales in perturbation error growth will be discussed.