Geoscience Reference
In-Depth Information
element temperature (water vapor) after convection. The KF scheme assumes that
at least 90 % of the environmental convective available potential energy (CAPE)
is consumed over
T
c
, which is limited between 1800 s and 3600 s (
Kain 2004
;
Kain and Fritsch 1993
).
T
c
is proportional to the model grid spacing and inversely
proportional to averaged winds between 500 hPa and the lifting condensation level.
The KF scheme is originally designed for a mesoscale model with grid spacing of
20-50 km, thus
T
c
matches well with the lifetime of a convective cell. This scheme
is dictated by the time it takes the cloud to grow to the point that precipitation forms
and the time it then takes the precipitation to fall to low levels (
Emanuel 1994
).
However, given typical mean horizontal wind speeds of
ms
1
,
T
c
will be fixed
to 1800 s if grid spacing is less than 18 km.
Narita and Ohmori
(
2007
)and
Saito
et al.
(
2007
) suggested that a shorter
10
T
c
of 900 s can improve the QPF with a 10 km
grid spacing in the Japan Meteorological Agency's operational mesoscale model.
The other parameter to be optimized,
, controls microphysical feedback from
the parameterized convection to its environment (
Correia et al. 2008
), and its math-
ematical formulation follows
Ogura and Cho
(
1973
) as function of the amount of
condensate at the layer bottom, the amount of condensate lost by the parameterized
updraft, the layer depth and updraft velocity.
The value of constant
c
s
1
in the old KF scheme (
Kain and Fritsch
c
is
0:01
s
1
in the new KF scheme of the WRF model.
2004
).
Correia et al.
(
2008
) found a smaller value of
1990
,
1993
), and is set to
0:03
s
1
directly increases
the hydrometeor feedback at the expense of the convective precipitation. Many
studies have shown the auto-conversion processes are responsible for determining
the organization and structure of convective systems (
Correia et al. 2008
;
Tao
et al. 1995
;
Zhu and Zhang 2004
).
0:005
27.2.2
Case Description and Experiments Design
Typhoon Rusa (2002) landed over the southwestern part of the Korean Peninsula
(KP) at 0600 UTC 31 August, 2002. Its central sea-level pressure stayed between
950 and 960 hPa. It moved across the Korea Peninsula and produced a 100-year
record-breaking heavy precipitation amount of
mm d
1
at Kangnung, located at
the central-eastern coast of the KP (
Gu et al. 2005
;
Lee and Choi 2010
;
Park and
Lee 2007
).
In this study, the numerical model WRF Version 3.2 is used. The computational
domain has a size of
870
;34
ı
N), with a grid
spacing of 25 km. The initial and boundary conditions are supplied by the NCEP
Final Analysis (FNL) data on
127
ı
E
1;800
km
2;100
km centered at (
1
ı
1
ı
with 6 h interval and the 3DVAR is used
to assimilate conventional surface and sounding observations. The simulations are
initiated at 00 UTC 30 August 2002, and ended at 12 UTC 31 August 2002. Schemes
for physical processes include: the YSU PBL, the WSM3 simple ice microphysics,
the Dudhia radiation and RRTM, and the Noah land surface model. For the CP
scheme, three groups of experiments are set up: (1) the KF scheme with default
Search WWH ::
Custom Search