Geoscience Reference
In-Depth Information
Fig. 27.1
The ETSs of
precipitation forecast in terms
of various thresholds (in mm)
of NOCP (
solid
), KFEX
(
dashed
)andOPTM
(
dash-dotted
) for a horizontal
resolution of 25 km. Numbers
beside each experiment name
are the sum of ETSs. The
precipitation is accumulated
from 1200 UTC 30 to
1200 UTC 31 August 2002
1
NOCP_2.33
KFEX_2.76
OPTM_7.84
grid size: 25 km
0.8
0.6
0.4
0.2
0
0
30
60
90
120
150
180
210
240
270
300
Threshold (mm)
observation in terms of both location and amount of precipitation (Fig.
27.2
d). Both
NOCP (Fig.
27.2
b) and KFEX (Fig.
27.2
c) have spurious local precipitation maxima
(
mm d
1
) at central-western coast of South Korea, where the observation
records only
> 200
mm d
1
(see Fig.
27.2
a). The spurious maximum of KFEX is
larger than that of NOCP, resulting in lower forecast skills of KFEX at thresholds
less than 210 mm (see Fig.
27.1
). Meanwhile, those false local maxima did not
appear in experiments with the optimized parameters.
50
-
100
27.5
Conclusions
In this study, optimal estimation of two parameters in the Kain-Fritsch convective
parameterization scheme is performed to improve the quantitative precipitation
forecast (QPF) for Typhoon Rusa (2002), which brought heavy rainfall in the Korean
Peninsula. The micro-GA is applied to find the best parameter values with a QPF
skill score as a fitness function, using the WRF model at a grid spacing of 25 km.
Among the two parameters, the auto-conversion rate
s
1
c
has a default value of
0:03
while the convective time scale
T
c
has a default range between 1800 s and 3600 s.
It turns out that, in order to produce the highest QPF skill at least for this
tropical cyclone case,
s
1
; thus the auto-conversion
is considered to be effectively turned off. The optimized
c
should be optimized to
0:0004
T
c
value is 1922 s. By
applying a set of two optimized parameters, the performance of WRF with a 25 km
resolution has been maximized in terms of the QPF skill for Typhoon Rusa (2002).
In this study, we have applied the micro-GA only to improve the QPF skill
of a tropical cyclone. However, other forecast aspects of tropical cyclones, such
as track and intensity, can be also improved via optimal parameter estimation by
defining different fitness function (e.g., squared error of track distance, mean sea
Search WWH ::
Custom Search