Diagnosis of operational model track forecast error for Hurricane Ike (2008)

The National Hurricane Center (NHC) issues track forecasts every six hours for all active tropical cyclones in its area of responsibility. These forecasts are made at projection times of 12, 24, 36, 48, 72, 96, and 120 hours. On 9 September 2008 track forecast errors from numerical model track guidance and the NHC's official forecast for Hurricane Ike increased substantially compared to errors from forecasts during the previous couple of days. This period of elevated track error occurred three to four days prior to Ike's Texas landfall early on 13 September. During this time, much of the track model guidance and the official NHC forecasts demonstrated a southward bias, taking the center of Ike to a landfall location along the central or southern Texas coast, instead of the upper Texas coast near Houston/Galveston where landfall ultimately occurred. The average error of the NHC 72-h track forecasts issued on 9 September was 92.8 nm, a 38% increase in the average error compared to the NHC 96-h track forecasts issued on the previous day (66.8 nm). Errors from many of the track models were even larger; the average 72-h forecast error for the interpolated GFS (GFSI) model on 9 September was 131.5 nm. This increase in track forecast error occurred during a critical time when federal, state, and local government agencies were making mitigation and disaster response decisions.

Examination of the 500-hPa height analyses from the members of the NCEP Global Ensemble Forecast System (GEFS) at 0000 UTC 9 September 2008 show significant spread in the strength of the mid-level subtropical ridge over the southern United States north of Ike. Ensemble members that showed a stronger (weaker) ridge north of Ike during the model integration moved Ike on a track farther to the left (right) of the track of the GEFS ensemble mean, which was close to the actual track of Ike. Given these observations, an experiment was performed where the initial conditions in the GFS run from 0000 UTC 9 September were altered to slightly weaken the ridge north of Ike. Results from this experiment showed Ike moving significantly to the right of the operational GFS forecast from that time and closer to the observed track of the hurricane. Additional experiments perturbing other synoptic scale features possibly impacting the track of Ike had limited impact on the hurricane's track. These results suggest that the evolution of the subtropical ridge was critical to determining Ike's ultimate track. This case also represents an instance where the track forecast at 4 and 5 days from the single-model GEFS was better than the typically superior multi-model consensus.