| description abstract | round 26 May 2008 a pronounced potential vorticity (PV) streamer penetrated from the North Atlantic into the western Mediterranean Sea followed by widespread dust mobilization over the Maghreb region of northwest Africa and a subsequent northward transport into central Europe. At the same time, strong southerly flow over the Mediterranean Sea caused heavy precipitation and flooding at the windward side of the European Alps. Using continuous and feature-based error measures, as well as ensemble correlation techniques, this study investigates the forecast quality and predictability of synoptic and mesoscale aspects of this high-impact event in operational ensemble predictions from nine meteorological centers participating in The Observing System Research and Predictability Experiment (THORPEX) Interactive Grand Global Ensemble (TIGGE) project. TIGGE is a recently established program providing ensemble forecasts in a standardized format, which allows for an exciting new multimodel approach to investigating the predictability of, for example, high-impact weather and its dynamics. The main conclusions from this study are that 1) the quality of the PV streamer forecasts degrades with lead time showing a general tendency toward too weak Rossby wave; 2) when focusing on the region around the streamer, most models show root-mean-square errors of the same magnitude or larger than the ensemble spread (underdispersive behavior); 3) errors are reduced by about 50% if the comparison is made to each center?s own analysis instead of the ECMWF analysis; 4) peak wind speeds over the Sahara tend to be underpredicted, with differences in model formulation dominating over differences in the representation of the PV streamer; and 5) ensemble-mean multimodel forecasts of 4-day accumulated precipitation appear accurate enough for a successful severe-weather warning. | |
