A set of criteria is developed to identify tropical cyclone (TC) formations in the Navy Operational Global Atmospheric Prediction System (NOGAPS) analyses and forecast fields. Then the NOGAPS forecasts of TC formations from 1997 to 1999 are verified relative to a formation time defined to be the first warning issued by the Joint Typhoon Warning Center. During these three years, the spatial distributions of TC formations were strongly affected by an El Niño-Southern Oscillation event. The successful NOGAPS predictions of formation within a maximum separation threshold of 4° latitude are about 70%-80% for 24-h forecasts, and drop to about 20%-30% for 120-h forecasts. The success rate is higher for formations in the South China Sea and between 160°E and 180° but is generally lower between 120° and 160°E. The composite 850-hPa large-scale flow for the formations between 120° and 160°E is similar to a monsoon confluence region with marked cross-equatorial flow. Therefore, it is concluded that the skill of NOGAPS in predicting TC formations with a monsoon confluence region pattern is lower than for other formation patterns. The number of false alarms (FAs) in NOGAPS is also examined. All the vortices in the 24-120-h forecasts that satisfy the above-defined criteria of a TC formation are identified and then compared with the formations (or nonformations) in the verifying analyses. The number of FAs is relatively low in the 24-h forecasts, increases through the 48-h forecasts, and then is about the same in the subsequent forecast times up to 120 h. Both the longitudinal distributions of the FAs during the three years and the meridional variations of the FA locations within a season are generally similar to the observed TC formations. Several large-scale features are compared for the successful and failed predictions during the 3-yr period. In general, the failed predictions have a slightly larger westerly vertical wind shear south of the TC formation location than in the successful predictions. The predicted relative humidity of the failed predictions is lower than in the successful formation forecasts. Thus, predicting too large a vertical wind shear and too dry an environment may be partially responsible for the failed cases. To further diagnose the failed predictions, various terms contributing to the angular momentum tendency are examined.
|Number of pages||21|
|Journal||Weather and Forecasting|
|Publication status||Published - Aug 2002|