25th Conference on Hurricanes and Tropical Meteorology

Monday, 29 April 2002: 5:00 PM
Performance of the NCEP Seasonal Forecast Model over the Atlantic tropical cyclone prone region
Wilbur Y. Chen, NOAA/NWS/NCEP/CPC, Camp Spring, MD
Poster PDF (670.6 kB)
The NCEP Seasonal Forecast Model (SFM) is already operational. By the fifth of every new month, the SFM will have generated an ensemble of 20 forecasts as well as an ensemble of 10 hindcasts, each out to 7 months into the future. The hindcasts cover 21 years, from 1979 to 1999, and therefore sufficient to provide a climatology which the forecast of climate anomalies can be based on.

This report assesses the performance of the SFM over the Atlantic tropical cyclone prone region. The model simulates well the tropical upper- and lower-level circulations. The vertical wind shear, which is known to be an important environmental factor that influences the cyclone activity, is therefore simulated well also. Besides climatology, the interannual variability of the wind shear field is also well simulated.

For the ASO peak cyclone activity seasons, the AC (Anomaly Correlation) skill scores during 1979 to 1999 yield an average value of 0.55 over the Atlantic cyclone prone region. This level of skill is remarkable when compared with an average of 0.3 for an extratropical hindcast. The one-month lead hindcasts, for 1994 and 1997 ASO cyclone season, foretell an observed change in wind shear anomaly from that of years of 1995, 1996, 1998 and 1999. The latter group of years was associated with an active cyclone season while the former group (1994 & 1997) a very subdued season.

The forecasts of vertical wind shear initiated in early July 2001 provide an opportunity to test the SFM's capability in forecast mode. The verifications for the forecasts are now available. The one-month lead seasonal forecast is found well validated. Furthermore, the multiple months long-lead monthly forecasts (for example, from earth July to September forecast) turn out to be also fairly successful.

The potential for long-lead seasonal and monthly forecast of the climatic conditions appears to be present. Further efforts to exploit the SFM's capability appears to be in good order.

Supplementary URL: