Geoscience Reference
In-Depth Information
In addition to the 100 m grid simulation, a series of experiments
examining the sensitivity to horizontal grid spacing in the range of cloud-
resolving simulations are performed. We set ∆
x
= 250 m, 500 m, 1 km,
2 km, and 4 km. These simulations are conducted in a larger computational
domain of 80 km
11 km, since a coarser-grid simulation will
require a larger domain for resolving cloud-scales, i.e. a couple of kilometers.
We have confirmed that the difference in computational area does not
affect significantly the convection and transport processes by comparing
the results with the two different computational areas in the case of 250 m
grid. In addition, coarser vertical grid spacings of 20-650 m (36 levels) are
used. In order to enhance the vertical mixing with these vertical grids, a
type of nonlocal mixing scheme
19
that modify the turbulence length-scale
in the vertical is employed.
The initial base state is set to the horizontal averages over the
computational area after a three-day spin-up run with the 250-m grid
starting with the vertical profile of Yinchuan, China, in the southern Gobi
Desert, at 0600 LT (local time at this longitude) on 13 April 2002. After
this spin-up run, a clear diurnal variation was represented. This base state,
shown in Fig. 1, is used for initializing the model with random temperature
perturbations added below the 1 km height, and has been used in our
previous studies.
6, 7
The model is integrated in time for 12 h for all the
simulation cases. The analyses are conducted for the model outputs at
300-s interval.
20 km
×
×
11
11
11
10
10
10
9
9
9
8
8
8
7
7
7
6
6
6
5
5
5
4
4
4
3
3
3
2
2
2
1
1
1
0
0
0
300
310
320
330
0.0
0.001
0.002
0.003
0
5
10
15
20
25
θ
(K)
(a)
Qv (kg/kg)
(b)
U (m/s)
(c)
Fig. 1. Initial vertical profiles of (a) potential temperature (in K); (b) water vapor
mixing ratio (in kg kg
−
1
); and (c) horizontal wind speed (in m s
−
1
,
x
-component).
