In this study, the authors study the prediction of an extra-tropical cyclone captured during the North Pacific Experiment (NORPEX). The cyclone was initialized in East-China and developed over Northwest Pacific from 1200 UTC 19 to 0000 UTC 24 February 1998. Penn State/NCAR MM5 is used to predicted the cyclone. MM5 4DVAR system is used to assimilate the GMS water vapor winds and total ozone mapping spectrometer (TOMS) data into the model initial conditions. The major results from the numerical experiments are as follows:
1) The skill of the model forecast degraded as the forecast duration was lengthened. The model was able to capture the cyclongenesis if NCEP analysis was used as initial conditions at 1200 UTC 20 February 1998. However, it suffered a significant degradation and failed to predict a major cyclone over the Northwest Pacific if the NCEP analysis at 1200 UTC 19 February 1999 was used as initial conditions.
2) Initial conditions incorporating the 6-hour GMS water vapor winds and TOMS data from 1200 UTC to 1800 UTC February 1998 using the MM5 4DVAR system were shown to improve forecasts of cyclongenesis substantially.
3) The improvement of forecasts of cyclogenesis is sensitive to the model resolutions for both assimilation and prediction. There is a best resolution for assimilating the GMS water vapor winds and TOMS data, which well reproduce the intensity of the cyclone.
![](images/flag.png)
![](images/blue_ribbon.png)