5.1 General remarks on the data format

We have tried for consistency reasons to keep as close as possible the same data format as last year for the ASTEX Stratocumulus case. However, since the different nature of the type of clouds we will require also different output. Specific emphasis will be put on the conditionally sampled fields as required in data sets D-G. Please precede each data set below by a line of up to 80 characters including the letter identifying the data set, the middle of the averaging hour, the investigator first initial_last name and, optionally, any further distinguishing characteristics, separated by spaces. If a variable is not available insert -99.9. If profiles are requested, start printing the highest level first. A first line of a dataset may look for instance

• A 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set A...
• B 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set B...
• C 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set C...
• D 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set D...
• E 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set E...
• F 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set F...
• G 2.5 P_Siebesma 3D run in 6.4km domain
  ...data for set G...
• H P_Siebesma 3D run in 6.4km domain
  ...data for set H...

Since models change, we are asking all participants to again fill out a model description section as was done last time. Since the purpose of an intercomparison is to understand how and why different models yield different results when applied to the same problem, please complete applicable parts of this section carefully.

5.2 Averaging procedures of the conditional sampled fields

For the sets D-G we require several conditionally sampled fields/fluxes. This output is of special interest for testing parametric assumptions of cloud parameterizations that use a mass flux approach. We require two types of decompositions:

• cloud decompositions: A grid point is diagnosed as cloudy if it contains liquid water.

• cloud core decompositions: A grid point is diagnosed as a cloud core point if it contains liquid water and if it is positively buoyant.

theta_v = theta * ( 1 + 0.61 * q_t - 1.61 * q_l )

In order to make the averaging procedure of the conditional sampled fields/fluxes more clear, we introduce some notation:

N = number of grid points at height z. (i.e N = 64*64 for the 3d models)
T = number of time steps in one hour.
M = N*T
 
Cloud indicator function I_cl(x,y,z,t) = 1 if grid point is cloudy
		         I_cl(x,y,z,t) = 0 otherwise

<f(z)>_cl =   sum   f(x,y,z,t) * I_cl(x,y,z,t) /  sum   I_cl(x,y,z,t)
	    (x,y,t)			        (x,y,t)

<a(z)>_cl =   sum   I_cl(x,y,z,t) /  M
	    (x,y,t)

<a(z)*f(z)>_cl=   sum   f(x,y,z,t) * I_cl(x,y,z,t) /  M
	         (x,y,t)

I_co(x,y,z,t) = 1    if the grid point is cloudy and positively buoyant
I_co(x,y,z,t) = 0    otherwise

The vertical integrated fields (see Set H) are defined as

 {f} = sum <f(z)> * rho(z) * dz
 	z

Primed quantities always mean deviations from the spatial horizontal slab average

5.3 Required output for 2d and 3d models

Sets A, B and C compare the time evolution of various variables through the last three hours of the simulation period. Use for these sets hourly (2-3(, (3-4), (4-5) and (5-6) average values and enter for the 'simulation time' the midpoint of the hour in the header (see above).

Set A compares the horizontal mean profiles averaged over the last four hours (2-3), (3-4), (4-5) and (5-6) of the simulation:

1. height (in meters), where the following mean values locate,
2. <U> (in m/s)
3. <V> (in m/s)
4. potential temperature <theta> (in K)
5. water vapour mixing ratio <q_v> (g/kg)
6. liquid water mixing ratio <q_l> (g/kg)
7. fraction of cloudy grid points <a>_cl (0-1)
8. reference density profile rho(z) (kg/m^3) (use mean density for compressible models)

Set B compares the hourly average (2-3), (3-4), (4-5) and (5-6) vertical profiles of turbulent velocity variances, skewness, and vertical turbulent fluxes (resolved plus subgrid, except where explicitly mentioned):

1. height (in meters), where the following fluxes in B locate,
2. <u'^2 + v'^2> (in m^2/s^2)
3. <w'^2> (in m^2/s^2)
4. Skewness profile <w'^3> / <w'^2>^(3/2)
5. x-momentum flux <u'w'> (in 10^-2m^2/s^2)
6. y-momentum flux <v'w'> (in 10^-2m^2/s^2)
7. liquid water potential temperature flux <w'thetal'> (10^-2 K m/s)
8. subgrid liquid water potential temperature flux (10^-2 K m/s)
9. total water flux <w'qt'> (10^-5 m/s)
10. subgrid total water flux (10^-5 m/s)
11. liquid water flux <w'ql'> (10^-5 m/s))
    
12. virtual potential temperature flux <w'theta_v'> (10^-2 K m/s))
    

Set C compares the hourly averaged (2-3), (3-4), (4-5) and (5-6) vertical profiles of the resolved turbulent kinetic energy (in m^2/s^2) and its budget terms (in 10^-4 m^2/s^3):

1. height (in meters), where the following variables locate,
2. resolved turbulent kinetic energy <q'> = <0.5*(u'^2 + v'^2 +w'^2)> where u' = u - <u> etc.
3. resolved shear production - (<u'w'>d<u>/dz + <v'w'>d<v>/dz)
4. resolved buoyancy production <w'b'>
5. resolved transport (turbulent plus pressure transport)
   -d/dz<w'(q'+p'/rho(z)) -u'*tau_13 - v'*tau_23 - w'*tau_33>
6. dissipation <tau_ij*du_i/dx_j> summed over i, j = 1,2,3.
   Here u_i are the velocity components and tau_ij are the components of the subgridscale stress tensor.
7. storage (i. e. (<resolved TKE>hour)- < resolved TKE>hour-1))/3600s) (in 10^-4 m^2/s^2)
8. residual (3) + (4) + (5) - (6) - (7)

The formulas above are correct for a Boussinesq fluid, but should be modified as needed to correspond to the equation set you are using.

Set D compares the hourly averaged (2-3), (3-4), (4-5) and (5-6) vertical profiles of conditionally sampled cloud fields.

Remark: The vertical velocity has to be taken relative to the specified large-scale subsidence so that <w>=0.

1. height (in meters), where the following variables locate,
2. fraction of cloudy grid points <a>_cl (0-1)
3. average over all cloudy grid points of vertical velocity <w>_cl (m/s)
4. average over all cloudy grid points of liquid water potential temperature (K)
5. average over all cloudy grid points of total water content (g/kg)
6. average over all cloudy grid points of liquid water content (g/kg)
7. average over all cloudy grid points of virtual potential temperature (K)

Set E compares the hourly averaged (2-3), (3-4), (4-5) and (5-6) covariances <wx>_cl for w and x=(theta_l, theta_v, q_t, q_l) for the conditionally sampled cloudy grid points.

1. height (in meters), where the following variables locate,
2. <a * w>_cl (m/s)
3. <a * wtheta_l>_cl (K m/s)
4. <a * wq_t>_cl (g/kg m/s)
5. <a * wq_l>_cl (g/kg m/s)
6. <a * wtheta_v>_cl (K m/s)

Set F compares the hourly averaged (2-3), (3-4), (4-5) and (5-6) vertical profiles of conditionally sampled cloudcore fields.

Remark: The vertical velocity has to be taken relative to the specified large-scale subsidence so that <w>=0.

1. height (in meters), where the following variables locate,
2. fraction of cloudcore grid points <a>_core (0-1)
3. average over all cloudcore grid points of vertical velocity <w>_core (m/s)
4. average over all cloudcore grid points of liquid water potential temperature (K)
5. average over all cloudcore grid points of total water content (g/kg)
6. average over all cloudcore grid points of liquid water content (g/kg)
7. average over all cloudcore grid points of virtual potential temperature (K)

Set G compares the hourly averaged (2-3), (3-4), (4-5) and (5-6) covariances <wx>_core for w and x=(theta_l, theta_v, q_t, q_l) for the conditionally sampled cloudcore grid points.

1. height (in meters), where the following variables locate,
2. <a * w>_core (m/s)
3. <a * wtheta_l>_core (K m/s)
4. <a * wq_t>_core (g/kg m/s)
5. <a * wq_l>_core (g/kg m/s)
6. <a * wtheta_v>_core (K m/s)

Set H compares the time evolution throughout the last four hours of cloud cover, vertical integrated liquid water and TKE.

Take instantaneous values with a time interval of one minute. See ( see section 5.2 ) for precise definition of vertical integration

1. Simulation time (minutes)
2. Fractional cloud cover (in percent), defined as the fraction of grid columns that contain cloud water.
3. Vertical Integrated liquid water (in kg/m^2)
4. Vertical Integrated turbulent kinetic energy (in kg/s^2)

Set I compares hourly averaged "in_cloud" variances of some fields x= { theta_l, theta_v, q_t, q_l }.

These are defined as: <sig(x)>_cl = < (x - <x>_cl)^2 >_cl

1. height (in meters), where the following variables locate,
2. <sig(theta_l>_cl (K^2)
3. <sig(q_t>_cl (g/kg)^2
4. <sig(q_l>_cl (g/kg)^2
5. <sig(theta_v>_cl (K^2)

5.4 Required output for 1d models

Not all the output requested in section 5-3 is feasible for 1d-models. We therefore require here a special output set for the 1d models which is a small subset of the requested output for the 2d and 3d models. We request 4-hourly averaged values over the whole period of 36 hours. (This means 9 sets of profiles for each dataset)

Set A compares the 4-hourly averaged profiles averaged:

1. height (in meters), where the following mean values locate,
2. potential temperature <theta> (in K)
3. water vapour mixing ratio <qv> (g/kg)
4. liquid water mixing ratio <ql> (g/kg)
5. cloud fraction (0-1)
6. reference density profile rho(z) (kg/m^3)

Set B compares the 4-hourly averaged fluxes

1. height (in meters), where the following fluxes in B locate,
2. liquid water potential temperature flux (10^-2 K m/s)
3. total water flux (10^-5 m/s)
4. liquid water flux (10^-5 m/s))
   
5. virtual potential temperature flux (10^-2 K m/s))
   

Set D compares 4-hourly averaged other fields.(Relevant for evaluating the cloud parameterizations) Only for those parameterizations that can obtain these profiles)

1. height (in meters), where the following variables locate,
2. cloud cover (0-1)
3. convective mass flux (kg m^-2 s^-1)
4. incloud liquid water potential temperature (K)
5. incloud total water content (g/kg)
6. incloud liquid water content(g/kg)
7. incloud virtual potential temperature (K)
8. total turbulent kinetic energy (m^2/s^2)

Set H compares the time evolution of the fractional cloud cover and the vertical integrated liquid water. Take instanteneous values each 5 minutes.

Remark: Use maximum overlap to obtain the total cloudcover

1. Simulation time (minutes)
2. Fractional cloud cover (in percent)
3. Vertical integrated liquid water (in kg/m^2)

Set I compares the 4-hourly averaged tendencies of the various processes for theta and qv.

1. height (in meters), where the following variables locate,
2. dtheta/dt due to PBL-schene (K/day)
3. dtheta/dt due to cloud-scheme (K/day)
4. dtheta/dt due to large-scale forcing (K/day)
5. dqv/dt due to PBL-schene (g/kg)/day)
6. dqv/dt due to cloud-scheme (g/kg)/day)
7. dqv/dt due to large-scale forcing (g/kg)/day)