Geoscience Reference
In-Depth Information
Fig. 5: Estimated intensity error: (a-c) minimum central sea level pressure error (hPa)
and (d-f) maximum sustainable wind (m/s) as estimated in CNTL and DA experiments
w. r. t. best-fit observation from IMD with different initial conditions.
noticed that the DA experiments provides less error than that of the CNTL
simulations. The improvement of about 10-25% is seen in DA experiments
with respect to MSLP and MSW error.
5. Conclusions
From the present study, following few broad conclusions are drawn.
The performance of the mesoscale model is highly dependable on the global
model products. Thus the improvement in prediction of extreme weather events
like tropical cyclone can be achieved with more appropriate initial condition
to the model integration. It may be noticed that the data assimilation experiments
show significant improvement at the model initial condition as well as in the
subsequent forecast time. With the more realistic initial condition, the initial
vortex position of cyclones Aila and Gonu are improved for about 37% and
26% respectively. The track simulation is also improved in the subsequent
forecast time with about 30-40%. Marginal improvement is noticed in the
intensity prediction for both the cyclones, and further improvement is noticed
in subsequent forecast time.
Acknowledgement
The Indian National Centre for Ocean Information Services (INCOIS) is
gratefully acknowledged for providing financial support to carry out this
research. The authors owe thanks to NCEP/NOAA for providing WRF-NMM
model and GFS data for carrying out this study. The authors are also grateful
to India Meteorological Department for providing best track data of tropical
cyclones used for the validation of model results.
Search WWH ::




Custom Search