1 The ideal em_les test case:
3 This test case produces a large-eddy simulation (LES) of a free
4 convective boundary layer (CBL). The environmental wind (or
5 the initial wind profile) is set to zero in this default case.
6 The turbulence of the free CBL is driven/maintained by the surface
7 heat flux, which is specified in the namelist as tke_heat_flux=0.24
10 A random perturbation is imposed initially on the mean temperature
11 field at the lowest four grid levels to kick off the
12 turbulent motion. Double periodic boundary condition
15 The default version uses a grid resolution of
16 dx=dy=100m and dz=50m, which is considered to be rather coarse
17 for an LES of the CBL. A typical grid mesh for an LES
18 of the CBL is dx=dy=50m and dz=20m. An LES flow field is more
19 accurate when the grid resolution is finer because it resolves
20 more turbulent scales.
22 This LES version uses the Deardorff's TKE scheme to compute the SGS
23 eddy viscosity and eddy diffusivity for turbulent mixing, that is
24 diff_opt=2 and km_opt = 2 in the namelist. The Coriolis parameter
25 is set to f = 10^-4/s.
27 It takes at least 30 minutes of simulation time to spin up the
28 turbulent flow field; only after the spin-up, the turbulence inside
29 the CBL is considered well established. A sign of well-established
30 turbulence is that the total (i.e., the resolved-scale plus the
31 subgrid-scale) heat flux profile should decrease linearly with height
34 To simulate a CBL with a mean wind, change the initial wind
35 profile in the input sounding.
36 When pert_coriolis= true is set in the namelist, there is no need to
37 include the geostrophic wind terms in the right-hand sides of
38 the u and v equations for LESs with non-zero geostrophic wind.
40 Note, parameterization constants, c_s and c_k in this namelist are
41 different from the defaults and are the ones recommended to use with LES.
43 =========================================================================
44 Shallow convection case (added in V3.7)
45 Copy namelist.input_shalconv to namelist.input
46 Copy input_sounding_shalconv to input_sounding
48 As LES case but larger domain (100x100), top still at 2 km.
49 Differences are sounding 10 K cooler, geostrophic u-wind 10 m/s, isfflx=1 using
50 input_sounding's surface theta = 295 K as SST, water surface, monotonic
51 moist and scalar advection, mp_physics =6, no other physics except sf_sfclay_physics=1.
52 This produces a shallow cloud layer capped by a strong inversion that
53 grows in response to warmer SST.