About KNMI Colloquia and workshops Staff Publications Contact Sitemap Homepage Research KODAC: KNMI Operational Data Centre
Set up 3D models
General remarks
We ask for two simulations, one with and one without microphysics. For those of you with very limited computer resources the simulation with microphysics is allowed to start at t = 8 hours of the simulations without microphysics.
Domain Parameters and Boundary Conditions
  • Duration: 24 hours
  • Domain Size: 12.8 * 12.8 * 4.0 km
  • Number of Grid Points: nx = ny = 128, nz = 100
  • Implying a Resolution: dx = dy = 100 m, dz = 40 m
  • Boundary conditions:
    • Lateral: Periodic
    • Top : In order to minimize spurious reflection of upward propagating gravity waves, you may want to use a sponge layer for damping perturbations. The sponge layer should not start any lower than 200 m above the mean inversion height.
    • Bottom : Surface fluxes are parameterized. See below.

Wind and Thermodynamic Profiles
Based on the observed profiles from radiosondes and aircraft dropsondes during RICO in the period of December 16 - 2004 to January 8 - 2005, the following initial setup for the horizontal wind components (u,v), the liquid potential temperature (theta_l) and the specific total water content (q_t) is proposed. Other profiles such as pressure, absolute temperature, etc, can be deduced assuming hydrostatic equilibrium. Initially, it can be assumed that there is zero liquid water (q_l = 0.0), so that theta = theta_l and q_v = q_t. A table with the profiles for the prescribed vertical 40m resolution can be found in the Appendix 3d models.
  • u [m/s]
    z > 0    -9.9 + 2.0*10-3*z
  • v [m/s]
    z > 0     -3.8
  • q_t [g/kg]
    0 < z < 740   16.0 + (13.8 - 16.0) / (740) * z
    740 < z < 3260   13.8 + (2.4 - 13.8) / (3260 - 740)*(z - 740)
    z > 3260   2.4 + (1.8 - 2.4)/(4000 - 3260)*(z - 3260)
  • theta_l [K]
    0 < z < 740   297.9
    z > 740   297.9 + (317.0 - 297.9)/(4000 - 740) *(z - 740)
Surface Conditions
The surface fluxes are parameterized in the simulation, by using a prescribed sea surface temperature (SST) and prescribed drag coefficients C_m, C_h and C_q The SST is based on an average for the RICO composite period of 2004/12/16 - 2005/01/08.
  • SST = 299.8 K
  • C_m = 0.001229
  • C_h = 0.001094
  • C_q = 0.001133
  • wthetha_l = -C_h*|U|*(theta_l - SST*(p0/p)(Rd/cp) )
  • wqt = -C_q*|U|*(qt - q_sat(T_sfc) )
  • uw = -u*C_m*|U|
  • vw = -v*C_m*|U|
with |U| = (u2+v2)1/2 where the velocities (u and v) are the local values at the lowest grid point level in the model (similarly for theta_l and qt). The total momentum flux is equal to C_m*|U|2 = u*2.
Additional surface characteristics:
  • surface pressure: ps = 1015.4 mb
  • sea surface potential temperature (with reference pressure of 1000 mb): ths = 298.5 K
Large Scale Forcings and Radiation
The large scale advection and subsidence are based on the analysis of the RACMO HindCast for a 2 months period centered on the RICO Domain. The large scale forcings should be only applied on q_t, theta_l and not u and v.
  • Large Scale Subsidence w [m/s] Apply the subsidence on the prognostic fields of q_t, theta_l.
    0 < z < 2260     - (0.005/2260) * z
    z > 2260     - 0.005
  • Large Scale Horizontal Liq. Water Pot. Temperature Advection combined with Radiative Cooling [K/s]
    z > 0     -2.5 / 86400
  • Large Scale Horizontal Moisture Advection [(g/kg)/s]
    0 < z < 2980    -1.0 / 86400 + (1.3456/ 86400) * z / 2980
    z > 2980    4. *10-6
The Geostrophic Wind
The zonal u-component of the geostrophic wind is decreasing with 2.0 * 10-3 s-1 corresponding with the observed wind above the mixed layer. The geostrophic v-component is assumed to be equal to the meridional wind v.
u    z > 0    -9.9 + 2.0 * 10-3 * z
v    z > 0    -3.8
Initial pertubations and translation velocity
The 3d model is initialised with random fluctutions of theta_l and q_t given by:
theta_l    [-0.1 , +0.1 ] (K)
q_t    [-2.5*10^-2, +2.5*10^-2] (g/kg)
Initial subgrid profile of subgrid TKE:
TKE    z > 0    1 - z/4000 m2/s2
In order to minimize numerical errors associated with advection we propose to translate the model domain with -6 and -4 m/s in the x and y direction, resp.
Microphysics
Simulations with a fixed number of cloud droplets should use a cloud droplet concentration of: 70 * 106 [m^-3]. This number is based on an average of best estimates of the active cloud droplet number concentration of four (out of six) flights during the three week period (calculated by Frederic Burnet and Jean-Louis Brenguier).
People needing a CCN concentration for initialization should use: 100*106 [m^-3]. This number is loosely based on data supplied by Jim Hudson (1% supersaturation). In the case that your microphysical model needs other - here not specified - information, please contact Margreet van Zanten.
Other parameters
  • Latitude: 18.0 N Degr.
  • Longitude: 61.5 W Degr.
  • c_p: 1005. J kg^-1 K^-1
  • g: 9.81 m s^-2
  • Rd: 287. J kg^-1 K^-1
  • L: 2.5 * 10^6 J kg^-1
  • surface pressure: 1015.4 mb
Updates
  • Typo (2250 instead of 2260) corrected in subsidence expression. [Nov 14, thanks to A. Ackerman]
  • Height limits in specification theta_l and q_t profiles adjusted (now it is more clear that profiles are continuous)[Nov 15, thanks to A. Cheng]
  • Coefficient and height adjusted in specification of large scale horizontal moisture advection in order to make profile continuous with more numbers behind the dot.[Nov 15, thanks to A. Cheng]
  • Moisture value at highest level adjusted to 1.8 instead 1.6 in order to stay closer to stationarity for moisture at upper levels [Nov 22, thanks to A. Ackerman]

© KNMI