P2.2 Characteristics of the TIGGE multimodel ensemble prediction system in representing forecast variability associated with extratropical transition

Wednesday, 18 April 2012
Heritage Ballroom (Sawgrass Marriott)
Julia H. Keller, Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany; and S. C. Jones, J. L. Evans, and P. A. Harr

Several times per year tropical cyclones start to recurve after their typical tropical life-cycle. They start to interact with the mid-latitude flow and may be transformed into an extratropical system. Such an extratropical transition (ET) process of a tropical cyclone often leads to a reduction in the predictability for the synoptic development of the cyclone itself as well as for the downstream region. Recent studies investigated the predictability during ET events based on the variability among members of single operational medium-range ensemble forecasts. The new THORPEX Interactive Grand Global Ensemble (TIGGE) provides an opportunity to extend these previous studies. TIGGE was established in the world weather research program THORPEX and combines the forecasts of 10 different EPS, operated at weather services all over the world. They are based on different assumptions, initial perturbation methods, resolutions and they also differ in the number of ensemble members, contained in the forecast. Thus TIGGE offers the possibility to compare predictability during ET events in a number of different ensemble prediction systems (EPS). Furthermore, TIGGE may show new possible development scenarios that could not be gained, using one single EPS.

To extract the information, contained in the ensemble forecasts, we perform an empirical orthogonal function (EOF) analysis on the variance-covariance-matrix of the forecast field in question. By calculating the principal components we get information, how each member, contained in the ensemble forecast contributes to the obtained EOF distribution. Using a fuzzy clustering, all members that show a related contribution, are grouped together. Thus, this analysis process allows us to extract possible development scenarios out of the ensemble forecast and at the same time we gain information about the possibility of these scenarios. For our investigations eight of the ten TIGGE EPS are used and interpolated to the same horizontal resolution and pressure levels. The investigations are mainly done for the geopotential height at a 500 hPa pressure level.

Using this data base and analysis technique, we perform case studies of several tropical cyclones, which underwent ET in 2008. The focus here is to study how the TIGGE multi model EPS behave by applying our analysis technique thereon, what kind of different possible development scenarios can be found for the ET process and how the individual EPS, contained in TIGGE, contribute to these scenarios. 18-2011-->

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner