The forecasted tracks are sensitive to initialisation parameters, they depend on the procedure to split the analysis fields into vortex and environmental flow and on the method to initialize the asymmetries of the implanted synthetic vortex. Differences between the forecasted tracks and the NHC's "best track" can be attributed to some degree to changes in strength and orientation of the vortex-asymmetries during the forecast period. The application of time-dependent boundary conditions reduces the mean forecast error of some storms of the 1996 Atlantic hurricane season.
Sensitivity studies are conducted to assess the potential to create individual deep layer means fields, depending on the particular structure of the storm in question. Furthermore, the sensitivity of the forecasted tracks to data sets originating from different forecast centres, is investigated.