***************************************************************** Run Description and Setup for ASTEX Lagrangian GCSS Intercomparison This intercomparison is based on the hourly ASTEX Lagrangian dataset. Follow this link to read about this dataset and link to the data directory. All filenames herein are relative to that directory. (a) Full Lagrangian simulations We propose two simulations appropriate for 1D and 2D models that can be used in a 'Lagrangian' mode in which the properties of air above the boundary layer are relaxed to specified time-varying profiles: (1) Lagrangian 1 (42 hours, starting on June 12.67) (2) Lagrangian 2 (40 hours, starting on June 18.93). The initial conditions for simulation 1 are in the files lagr1/hourly/lagr1_00. The changing boundary conditions at hour hh (together with the observed boundary layer sounding) are in file lagr1_hh. See file README.hourly for a complete description of the hourly data, as well as a description of currently available verification datasets. (b) Short simulations We also propose two shorter simulations feasible for 3D LES models (and also interesting comparisons for 2D and 1D models). These are based on shorter intervals within the Lagrangian experiments. The simulated periods have been chosen so that upper air thermodynamic properties change negligibly during the integration, making a 'Lagrangian' simulation simpler to carry out: (3) A three hour long simulation starting at hour 12 of Lagrangian 1 (June 13.167) suitable for both 2D and 3D models. At this time, stratocumulus with sustained drizzle rates of more than 1 mm/day at the ground was observed as the SST under the air column increased very rapidly. This simulation is largely (but not entirely) nocturnal. The initial conditions for simulation (3) can be found in file lagr1/hourly/lagr1_12 and the variation of SST, lat, lon can be taken from file lagr1/lagr1/hourly/scalars. (4) A six hour long simulation suitable for both 2D and 3D models of a deep 'decoupled' boundary layer with cumulus clouds rising into patchy stratocumulus, starting at hour 32 of Lagrangian 2 (June 20.25) . The goal here is to compare the predictions of cloud fraction and optical properties by different models. The initial conditions for simulation (4) can be found in file lagr2/hourly/lagr_32 and the variation of SST, lat, lon can be taken from file lagr2/hourly/scalars. Simulations 3 and 4 have been chosen so that the time variations of the above-inversion conditions appear to be small over the length of the simulation. For simplicity, we will run them with all external boundary conditions fixed at their initial values in the hourly sounding archive except SST and the diurnal cycle. SST and solar zenith angle should be computed for these simulations as discussed in the Boundary Conditions section below. Both cases 3 and 4 include daytime periods, though for case 3, solar radiation may be comparatively unimportant. ---------------------------------------------------------------- Initial & boundary conditions, modelling assumptions and other useful information (a) Time interpolation For all simulations, we recommend that between hours, linear interpolation in time be used to get continuously varying boundary data. (b) Domain translation velocity Those of you with LES type models will probably choose to use a reference frame that moves along approximately with the mean flow to minimize numerical stability constraints on the model timestep. We see two choices for handling domain translation: 1. Choose a constant domain translation speed (in x, y directions). The frame does not accelerate relative to the earth's surface. 2. Choose a time-varying domain translation speed (in x, y directions) (so as to follow changes in the mean boundary layer wind with time in a Lagrangian simulation). The frame now accelerates relative to the surface of the earth, and the acceleration of the frame should be accounted for in the model momentum equations. (c) Initial conditions in saturated air As mentioned in README.hourly, the water vapor mixing ratio in the hourly sounding may not be saturated even though mean liquid water is present, due both to horizontal inhomogeneity and measurement errors. Thus, to ensure thermodynamic consistency with the measured mean liquid water in your initial conditions, you should replace the water vapor mixing ratio by the saturation water vapor mixing ratio at any heights in the sounding at which either (i) the droplet concentration is at least 0.5 as large as the estimated in-cloud droplet concentration or (ii) the liquid water mixing ratio ql is at least 0.05 g/kg. (d) Baroclinicity If the geostrophic wind varies with height, thermal wind balance implies that there must be horizontal temperature gradients, and it is inconsistent to assume that the temperature and pressure fields are periodic in the horizontal. However, if your model cannot handle this, you should ignore the mean horizontal temperature gradients, and treat temperature and pressure fields are periodic in the horizontal. This assumption, most accurate for a small model domain such as we are using, is akin to including mean subsidence without explicitly including the divergent component of the horizontal mean wind. Note how baroclinicity is handled in the Model Description section of your output in the Physical Parameterizations section. (e) Downwelling longwave radiation Boundary layer cloud simulations are well known to be quite sensitive to the downwelling flux of longwave radiation at the boundary layer top. To ensure we are all in the same ballpark, note that the hourly sounding files all contain a measured value of downwelling LW flux at 700 mb (Austin and Bretherton 1995). With the exception of June 19.4-19.75 during Lagrangian 2, during which there was cirrus cloud cover, these downwelling fluxes agree very well with clear air calculations based on temperature and moisture profiles taken from ECMWF analyses done for ASTEX (Austin and Bretherton 1995). The hourly soundings include coarse resolution ECMWF-analyzed profiles above 700 mb suitable for radiative transfer calculations. If your code uses a full atmospheric column radiation subroutine, use it but check that your downwelling longwave flux at 700 mb is close (within 10 W/m2) of the hourly specified value. If it isn't you might want to adjust your mixing ratio at pressures below 700 mb to match the desired downwelling longwave flux at 700 mb. If your code specifies the radiative fluxes at the top of the model domain, use the hourly value given, correcting it to the height of your domain top as described in README.hourly in the discussion of LW_dn in the header of the hourly file. (f) Relaxation in free atmosphere for Lagrangian runs For simulations (1) and (2) in which free atmospheric conditions are to be varied in a specified fashion with time, use the following relaxation, which has a time constant long enough to not suppress the transmission of gravity waves and which is only applied significantly above the maximum inversion height: zimax = max over all columns of local zi local zi = height at which gradient of total water mixing ratio is maximum. For any scalar field F, add a term onto dF/dt at heights z >= zimax + 75 m: (dF/dt)_relax = - (F - Fe(z,t))/tau where F is the actual value of the scalar, Fe(z,t) is its target value, and tau = 3 hrs. No relaxation is used on the velocities, because the time varying geostrophic wind maintains a reasonable wind profile. (g) Upper boundary condition for gravity waves In the effort to be as realistic as possible, you may want to use either a sponge layer for damping perturbation momentum or a radiation upper boundary condition to minimize spurious reflection of upward propagating gravity waves. The sponge layer should not start any lower than 200 m above the mean inversion height. (h) For use with your radiation scheme: Solar constant: 1323 W/m2 (due to the variation of the earth-sun distance d from its mean value dm, the solar ``constant'' is reduced from its mean value. For June 16, the ratio (dm/d)**2 = 0.9690, so S = smean * (dm/d)**2 = 1323 W / m^2.) If you need formulas for the variation of zenith angle with time, lat, lon: cosine(zenith angle) = max( -cos(sun_lat)*cos(sun_azim)*cos_lat + sin(sun_lat)*sin_lat, 0.) where lat = latitude (degrees N), sin_lat = sin(pi*lat/180.) cos_lat = cos(pi*lat/180.) lon = longitude (degrees E, <0 for degrees W) day = decimal day of June (UTC). month = 6 + day/30. pi = 3.1415926... sun_azim = 2*pi*(day - lon/360.) (radians) sun_lat = (pi*23./180.)*cos((month - 6.75)*pi/6) (radians) (h) Changing surface pressure Over the course of each Lagrangian, surface pressure changes slightly (by 2-4 mb over 40 hours). If you use reference profiles in your dynamical equations and to compute thermodynamic coefficients, it should suffice to choose reference values based on the initial sounding. However, be aware of the surface pressure variations in calculations (such as finding saturation mixing ratio and temperature knowing potential temperature and height) which depend upon the horizontal mean pressure at a given height at a given time. (i) Use your model's regular scheme for calculating surface heat, moisture and momentum fluxes, and for determining eddy diffusivity, TKE, etc. at the lowest gridpoint above the sea-surface. Contact one of us if you would like to know what we use. However, the purpose of this intercomparison is to compare the effect of different approaches to physical parameterization as well as different numerics. (j) Other Details * Set sea-surface albedo = 0.05 * Periodic lateral BC's * Mean subsidence specified from given horizontal divergence on hourly file header * Coriolis forcing based on given latitude (use fixed latitude of 35 degrees N if it is inconvenient to make f time-varying) * Holton's values for physical constants: g=9.81 m/s^2, R=287 J/(K-kg), Cp=1004 J/(K-kg), L=2.5e6 J/kg. If it is convenient for you to use these values, please do so. * Random Initialization : spatially uncorrelated random perturbation uniformly distributed between -0.1 and 0.1 K, applied to initial temperature field at all gridpoints below the initial inversion height. ----------------------------------------------------------------- Numerical Domain Specifications ********************************************* Case 1: Lagr1 ********************************************* Duration of run : 42 hours Initial sounding file : /lagr1/hourly/lagr1_00 Domain Height : 3000. m Domain Length (X) : 6400. m Vertical Grid Points : 60 (dz = 50 m) X Grid Points : 64 (dx = 100 m) ********************************************* Case 2: Lagr2 ********************************************* Duration of run : 40 hours Initial sounding file : /lagr2/hourly/lagr2_00 Domain Height : 3000. m Domain Length (X) : 6400. m Vertical Grid Points : 60 (dz = 50 m) X Grid Points : 64 (dx = 100 m) ********************************************* Case 3: Drizzle ********************************************* Duration of run : 180 min Initial sounding file : /lagr1/hourly/lagr1_12 Domain Height : 1500. m Domain Length (X) : 3200. m Domain Width (Y) : 3200. m Vertical Grid Points : 60 (dz = 25 m) X Grid Points : 64 (dx = 50 m) Y Grid Points : 64 (dy = 50 m) ********************************************* Case 4: Cumulus ********************************************* Duration of run : 360 min Initial sounding file : /lagr2/hourly/lagr2_32 Domain Height : 3000. m Domain Length (X) : 6400. m Domain Width (Y) : 6400. m Vertical Grid Points : 60 (dz = 50 m) X Grid Points : 64 (dx = 100 m) Y Grid Points : 64 (dy = 100 m) If your model uses variable vertical resolution, try to match the given resolution near the inversion. *************************************************************************** Data Format for Presenting Results Our intercomparison results will be presented in seven different sets, some split into two parts to avoid making lines of more than 80 characters that may be mangled by electronic mailers. The data format chosen here has been optimized for Steve Krueger's plotting programs. Sets F and G are optional, and investigate in more detail the TKE budget and use of conditional sampling. If you have a 1D model or if you don't wish to do parts F and G, just enter just the first five sets. The data format for presenting intercomparisons is modelled on Chin-Hoh's format for the first intercomparison, but because of the nature of the run, there are some differences, so read carefully. We have made an effort to only refer to quantitities (i. e., total liquid water, rather than cloud water separately) that also make sense for models with explicit microphysics. In what follows, f-bar refers to a horizontal average of f. The layer average of f is the vertical average of f-bar from the surface to the boundary layer top (inversion) height zi (which is defined in C1.2 below). Please precede each dataset below by a line of up to 80 characters including a comment identifier #k, the run number (1,2,3, or 4), letter identifying the dataset, (for sets D-G) the middle of the averaging hour, the investigator first initial_last name and, optionally, any further distinguishing characteristics, separated by spaces, e. g. #k 1 A1 S_Krueger 2D run in 6.4km domain ...data for set A1... #k 1 A2 S_Krueger 2D run in 6.4km domain ...data for set A2... #k 1 B S_Krueger 2D run in 6.4km domain ...data for set B... #k 1 C1 S_Krueger 2D run in 6.4km domain ...data for set C1... #k 1 C2 S_Krueger 2D run in 6.4km domain ...data for set C2... #k 1 D 0.5 S_Krueger 2D run in 6.4km domain ...data for set D... #k 1 E1 0.5 S_Krueger 2D run in 6.4km domain ...data for set E1... #k 1 E2 0.5 S_Krueger 2D run in 6.4km domain ...data for set E2... #k 1 F 0.5 S_Krueger 2D run in 6.4km domain ...data for set F... #k 1 G1 0.5 S_Krueger 2D run in 6.4km domain ...data for set G1... #k 1 G2 0.5 S_Krueger 2D run in 6.4km domain ...data for set G2... #k 1 D 1.5 S_Krueger 2D run in 6.4km domain ...data for set D... #k 1 E1 1.5 S_Krueger 2D run in 6.4km domain ...data for set E1... #k 1 E2 1.5 S_Krueger 2D run in 6.4km domain ...data for set E2... #k 1 F 1.5 S_Krueger 2D run in 6.4km domain ...data for set F... #k 1 G1 1.5 S_Krueger 2D run in 6.4km domain ...data for set G1... #k 1 G2 1.5 S_Krueger 2D run in 6.4km domain ...data for set G2... . . . 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. --------------------------------------------------------------------- Sets A1 and A2 document the initial conditions and attempt a crude radiative intercomparison between models. Set A1 (1) height (in meters), where the following initial values in A1/A2 locate, (2) U-bar (ground-relative, in m/s) (3) V-bar (ground-relative, in m/s) (4) potential temperature theta-bar (in K) (5) water vapor mixing ratio qv-bar (g/kg) (6) liquid water mixing ratio (including rainwater) ql-bar (g/kg) (7) Upward longwave radiative flux (W/m^2) (8) Downward longwave radiative flux (W/m^2) The data format for Set A1 is (f7.1,3f7.2,2f6.2,2f6.1) Set A2 (9) Upward shortwave radiative flux (W/m^2) for a zenith angle of 0 degrees (10)Downward shortwave radiative flux (W/m^2) for a zenith angle of 0 degrees (11)Upward shortwave radiative flux (W/m^2) for a zenith angle of 60 degrees (12)Downward shortwave radiative flux (W/m^2) for a zenith angle of 60 degrees (13)Initial reference density profile rho(z) (kg/m^3) (use mean density for compressible models) The data format for Set A2 is (4f7.1,f6.3) ---------------------------------------------------------------------- Sets B and C1/C2 compare the time evolution of various scalars through the whole simulation period. For set B, use instantaneous values with a sampling time interval of between 1 and 5 minutes for cases 3 and 4, and between 5 and 10 minutes for cases 1 and 2. (1) Simulation time (in min) (2) SST (K) (3) Cloud fraction (0-1), defined as the fraction of columns with at least one gridpoint in which liquid water mixing ratio exceeds 0.01 g/kg. (4) 'Satellite-detectable' cloud fraction (0-1), defined as the fraction of grid columns in which the column-integrated liquid water path exceeds 13 g/m^2 (this corresponds to a cloud optical depth of 2, roughly the minimum thickness of cloud detectable from its increased albedo compared to the sea-surface.) (5) Mean column-integrated liquid water path (in g/m^2). The mean is taken only over cloudy columns as defined in B.3. If there are no such columns, enter 0. (6) Mean of the cloud base height (defined as the lowest height at which the cloud fraction defined by D.7 exceeds 0.01) (in m). The mean is taken only over cloudy columns as defined in B.3. If there are no such columns, enter 0. (7) Domain-averaged precipitation rate at sea surface (in mm H20/day) (8) Maximum updraft strength in domain (in m/s) (9) Maximum liquid water (cloud water + rain water) mixing ratio in domain (in g/kg) The data format for Set B is (f7.1,8f9.3) ------------------------------------------------------------------------------ For sets C1/C2, use hourly average values. For the 'simulation time' enter the midpoint of the hour. While some of this information duplicates entries in Set D, this data format is more convenient for our plotting script. Other scalars duplicate set B. For these scalars, the hourly average in C is useful for comparing with other hourly average profiles and B helps pin down the variability of that scalar. Set C1 (1) Simulation time (in min) (2) Mean inversion height zi (in meters), defined as the horizontal average over all columns of the half grid level with the strongest decrease in total (liquid + vapor) water content with height (3) Cloud fraction (0-1), defined as the fraction of columns with at least one gridpoint in which liquid water mixing ratio exceeds 0.01 g/kg. (4) Mean column-integrated liquid water path (in g/m^2) (5) Domain-averaged precipitation rate at sea surface (in mm H20/day) (6) Surface sensible heat flux rho(0)*Cp*w'theta'-bar (in W/m^2) (7) Surface latent heat flux rho(0)*L*w'qv'-bar (in W/m^2) (8) Mean air temperature at first gridpoint above surface (in K) (9) Mean water vapor mixing ratio at first gridpoint above surface (in g/kg) The data format for Set C1 is (f7.1,8f9.3) Set C2 (10)Layer averaged total (resolved + subgrid) turbulent kinetic energy (in m^2/s^2) (11)Total solar + longwave radiative flux divergence between sea surface and 50 m above mean inversion height zi (in W/m^2) (12)Upward solar flux at top of numerical model domain (either 3000 m or 1500 m, depending on the case) (in W/m^2) (13)Downward solar flux at top of numerical model domain (in W/m^2) (14)Upward longwave flux at top of numerical model domain (in W/m^2) (15)Downward longwave flux at top of numerical model domain (in W/m^2) (16)Downward solar flux at sea surface (in W/m^2) (17)Downward longwave flux at sea surface (in W/m^2) The data format for Set C2 is (8f9.3) --------------------------------------------------------------------- Set D compares the horizontal mean profiles averaged over each hour of simulation (1) height (in meters), where the following mean values locate, (2) U-bar (in m/s) (3) V-bar (in m/s) (4) potential temperature theta-bar (in K) (5) water vapor mixing ratio qv-bar (g/kg) (6) liquid water mixing ratio (including rainwater) ql-bar (g/kg) (7) Cloud fraction: (fraction of saturated gridpoints at the given level, or area average of the grid point cloud cover if grid points can be partly cloudy) (0-1). (8) Reference density profile rho(z) (kg/m^3) (use mean density for compressible models) The data format for Set D is (3F8.2, 5F8.3). --------------------------------------------------------------------- Set E1/E2 compares the hourly averaged turbulent velocity variances, skewness, and flux profiles (resolved plus subgrid, except where explicitly mentioned): Set E1 (1) height (in meters), where the following fluxes in E1/E2 locate, (2) u'^2 + v'^2-bar (in m^2/s^2) (3) w'^2-bar (in m^2/s^2) (4) Skewness profile w'^3-bar/w'^2-bar^(3/2) (5) x-momentum flux u'w'-bar (in m^2/s^2) (6) y-momentum flux v'w'-bar (in m^2/s^2) The data format for Set E1 is (F7.1,2f6.3,f7.3,2F7.4) Set E2 (7) Heat flux rho(k)*Cp*w'theta'-bar (in W/m^2) (8) Water vapor flux rho(k)*L*w'qv'-bar (in W/m^2) (9) Convective liquid water flux rho(k)*L*w'ql'-bar (in W/m^2) (note that ql includes both cloud and rain water) (10) Water sedimentation (precipitation) flux rho(k)*L*sum over i of w_t(i)ql(i)-bar (in W/m^2), where w_t(i) is the terminal velocity and ql(i) is liquid water, summed over all categories i of liquid water. Sign convention is that w_t(i) 0. (11) Net (upward-downward) longwave radiative flux (W/m^2) (12) Net (upward-downward) shortwave radiative flux (W/m^2) The data format for Set E2 is (4f7.2,2f8.1). --------------------------------------------------------------------- Set F compares the hourly average vertical profiles of the resolved turbulent kinetic energy (in m^2/s^2) and its budget terms, each multiplied by 10**4 (in m^2/s^3): (1) height (in meters), where the following variables locate, (2) resolved turbulent kinetic energy q'-bar = 0.5*(u'^2 + v'^2 +0.5 m/s), a downdraft point (w -0.5 m/s), or an environment point (-0.5 .le. w .le. 0.5 m/s). Note that w is the vertical velocity relative to the specified large-scale subsidence, so that w-bar = 0. Set G1 (1) height (m) where the following values in G1/G2 are located (2) fraction of updraft grid points, sigma_u (0-1) (3) fraction of downdraft grid points, sigma_d (0-1) (4) average over all updraft points of vertical velocity, w_u (m/s) (5) average over all downdraft points of vertical velocity, w_d (m/s); note that sigma_u * w_u + sigma_d * w_d + sigma_e * w_e = w-bar = 0. (6) updraft mass flux, Mu= rho(z) * sigma_u * w_u (kg m^-2 s^-1) (7) downdraft mass flux, Md= rho(z) * sigma_d * w_d (kg m^-2 s^-1) The data format for set G1 is (f7.1,2f7.3,2f6.2,2f7.3) Set G2 (8) avg over all updraft points of virtual potential temperature, thv = theta ( 1. + 0.61 qv - ql ) (K) (9) avg over all downdraft points of thv (K) (10) avg over all points of thv, thv-bar (K) (11) avg over all updraft points of ql (liquid water mixing ratio including rainwater) (g/kg) (12) avg over all downdraft points of ql (g/kg) (13) avg over all points of ql, ql-bar (g/kg) (14) Updraft buoyancy flux rho(k)*cp*sigma_u*{average over all updraft points of w * thv'} (W/m^2), where thv' = thv - thv-bar. (15) Downdraft buoyancy flux rho(k)*cp*sigma_d*{average over all downdraft points of w * thv'} (W/m^2) The data format for set G2 is (8f8.3) ************************** MODEL DESCRIPTION ********************* | | | | Scientist Name: | Model Type (1D ensemble mean, 2D cloud resolving, or 3D LES): | -------------------------------------------------------------------- |(1) Computation | |Computer used for simulation: |CPU time: -------------------------------------------------------------------- -------------------------------------------------------------------- |(2) Numerical Domain | |Dimensions (fill in if different than specified in problem description): |Domain size in x-direction: |Domain size in y-direction: |Domain size in z-direction: |Number of grid points in x-direction: |Number of grid points in y-direction: |Number of grid points in z-direction: |Grid size in x-direction: |Grid size in y-direction: |Grid size in z-direction: | |Time step for dynamics: -------------------------------------------------------------------- -------------------------------------------------------------------- |(3) Numerical Technique | |Numerical method (finite-difference, spectral, etc.): |Advection scheme and its order of accuracy: |Time scheme and its order of accuracy: |Dynamical equations (elastic, anelastic, etc.): |Lateral boundary conditions: |Upper boundary condition: !How was translation velocity of the reference frame chosen? !Are accelerations due to time variation of the reference frame ! velocity compared to the ground included in the dynamical equations? -------------------------------------------------------------------- |(4) Physical Parameterizations | |Surface flux parameterization for heat, moisture, momentum: |Longwave radiation parameterization: |Shortwave radiation parameterization: |How were radiative fluxes above the computational domain handled? |Were horizontal temperature gradients in thermal wind balance | with the vertical shear of the geostrophic wind included? |Drizzle/Rain parameterization: |Other comments on physical parameterizations: -------------------------------------------------------------------- |(5) Turbulence Closure Scheme (if applicable) | |Turbulence closure (SGS or ensemble mean): |Specific turbulence closure type (Mellor-Yamada, etc.): |Variables predicted by the turbulence closure | (eddy diffusivities, turbulent kinetic energy, etc.): |Closure for dissipation rate: |Closure for turbulent dissipation length scale: |Closure for turbulent diffusion length scale: -------------------------------------------------------------------- -------------------------------------------------------------------- |(6) Documentation | |Please provide references that more fully describe your model, if |available. --------------------------------------------------------------------