Geoscience Reference
In-Depth Information
Fig. 2:
Difference in initial temperature, RH and winds between CTL and
3DVAR for Jal cyclone.
increases towards North Bay of Bengal. The model-simulated tracks for all 10
tropical cyclone cases (Table 1) in both experiments are compared with IMD
best track estimates (Fig. 3). The vector track picture for 10 cyclone cases
indicated 3DVAR experiments using conventional observations which reduced
the track position errors in a majority of simulated storms except for Sidar and
Aila. Improvement is noted in simulated vector track in 3DVAR experiments
from the model initial time till landfall of all cyclone cases except for Jal and
Sidr cases.
The forecast errors in various parameters for each case are calculated as
corresponding IMD estimates—simulated values for the parameters of central
pressure, winds and track positions. The mean errors in simulated track and
variation in mean errors of central sea level pressure (CSLP) and maximum
winds corresponding with IMD estimates in both experiments of 10 cyclone
cases are presented in Fig. 5. The mean position errors indicate considerable
decrease in initial position error. The track errors are reduced by 50 km at 24 h,
10 km at 48 h, 65 km at 72 h and about 100 km at 96 h integration (Fig. 4a)
with assimilation which shows substantial improvements in the model results
for simulated track with 3DVAR. The mean position errors indicate that there
is considerable decrease in initial position error.
The track errors are reduced by 25 km at 24 h, 8 km at 48 h, 34 km at 72 h
and about 4 km at 96 h integration (Fig. 4a) in 3DVAR experiments. This
indicates reduction of 11%, 13%, 3%, 16% and 1% in track errors at 12, 24,
48, 72 and 96 h of simulation period in 3DVAR runs. The time variation of
mean errors in CSLP indicated the control experiments give large errors of 9
hPa, 21 hPa, 18 hPa and 5 hPa at 24, 48, 72 and 96 h respectively (Fig. 4b) and
Search WWH ::
Custom Search