12B.5
Evaluation of dynamical track predictions for tropical cyclones in the Atlantic during 1995–98
David S. Brown, NPS, Monterey, CA; and M. A. Boothe, L. E. Carr, and R. L. Elsberry
Carr and Elsberry (1999; NPS Tech. Report) have described eight conceptual models that explain most cases of large (> 300 n mi at 72 h) track errors by the Navy Operational Global Atmospheric Prediction System (NOGAPS) and the Geophysical Fluid Dynamics Lab (Navy version – GFDN) models in the western North Pacific. Dunnavan et al. (1999; preprint in this conference) describes these same conceptual models applied during the 1998 season and for the UK Meteorological Office (UKMO) model. This study changes the focus to the Atlantic basin, with the hypothesis that the three dynamical models have similar characteristics and that cases with large errors in tropical cyclone (TC) track forecasts may be explained with the same eight conceptual models and with similar relative frequencies.
This study examines large errors made by the dynamical models for four seasons of Atlantic TC tracks (1995-98). The minimum threshold for a model error to be defined as “large” in the Atlantic is reduced from 300 n mi in the western North Pacific to 250 n mi to reflect the lower average model errors in the Atlantic. At this time, only the NOGAPS tracks and fields are available for all four years. NPS has access to the UKMO fields of only the last two seasons, and the GFDL fields are available for only the 1997 season. During the 1998 season, the Aviation model, which is the background field for the GFDL regional model, affected GFDL forecasts in a detrimental way. A fire at the National Center for Environment Prediction has so far prevented data acquisition for the GFDL for the earlier seasons of 1995-96.
Initial results suggest that the same model error traits that apply for TCs in other basins also affect the dynamical model tracks of Atlantic TCs. All three models show a tendency to over-represent TCs and other circulations rotating about or even merging with each other via the Excessive Direct Cyclone Interaction (E-DCI) process, just as was found in the western North Pacific. More complete results will be available for the preprint and presentation at the conference.
Session 12B, Tropical cyclone track forecasting I (Parallel with Sessions 12A and 12C)
Friday, 26 May 2000, 8:00 AM-10:00 AM
Previous paper Next paper