https://pcmdi.llnl.gov UCLA AGCM6.4 (4x5 L15) 1992, except that gravity-wave drag is not implemented and the prognostic depth of the planetary boundary layer (PBL) is not smoothed. The distinguishing feature of the model is that its code is written explicitly for massively parallel processing (MPP) computers. The model has generated 20 realizations of the AMIP experiment from different initial conditions. UCLA AGCM6.4 (4x5 L15) 1992, key documentation is provided by Arakawa (1972) , Arakawa and Lamb (1977 , 1981 ), Arakawa and Schubert (1974) , Arakawa and Suarez (1983) , Lord (1978) , Lord and Arakawa (1980) , Lord et al. (1982) , Randall et al. (1985) , Suarez et al. (1983) , and Takano and Wurtele (1982) . In addition, the parallelism and computational performance of the LLNL model are discussed by Wehner et al. (1995) and Mirin and Wehner (1995), while an initial validation of the model is performed by Wehner and Covey (1995). Issues relevant to coupling the parallel LLNL atmospheric model to an ocean model are discussed by Wehner et al. (1994). , 1981 ). The horizontal advection of momentum is treated by the potential-enstrophy conserving scheme of Arakawa and Lamb (1981) , modified to give fourth-order accuracy for the advection of potential vorticity (cf. Takano and Wurtele 1981 ). The horizontal advection scheme is also fourth-order for potential temperature (conserving the global mass integral of its square), and for water vapor and prognostic ozone (see Chemistry). The differencing of the continuity equation and the pressure gradient force is of second-order accuracy. Vertical Representation, Vertical Resolution, and Planetary Boundary Layer.
- Finite differences in modified sigma coordinates. For P the pressure at a given level, PT = 1 hPa the constant pressure at the model top, PI = 100 hPa the pressure at a level near the tropopause, PB the pressure at the top of the planetary boundary layer, and PS the pressure at the surface, sigma = (P - PI)/(PI - PT) for PI >= P >= PT (in the stratosphere); sigma = (P - PI)/(PB - PI) for PB >= P >= PI (in the troposphere above the PBL); and sigma = 1 + (P - PB)/(PS - PB) for PS >= P >= PB (in the PBL). The sigma levels above 100 hPa are evenly spaced in the logarithm of pressure (cf. Suarez et al. 1983
- The vertical differencing scheme after Arakawa and Suarez (1983)  and Tokioka (1978)  conserves global mass integrals of potential temperature and total potential plus kinetic energy for frictionless adiabatic flow. See also Vertical Resolution and Planetary Boundary Layer.
- Nonlinear second-order horizontal diffusion after Smagorinsky (1963)
 is applied (with a small coefficient) only to the momentum equation on the modified sigma levels (see Vertical Representation).
- Vertical diffusion is not explicitly included; however, momentum is redistributed by cumulus convection (see Convection).
- Shortwave radiation is parameterized after Katayama (1972)
 with modifications by Schlesinger (1976)
. Following Joseph (1970)
, the radiation is divided into absorbed and scattered components. Absorption by water vapor and ozone is modeled using wavelength-integrated transmission functions of Yamamoto (1962)
 and Elsasser (1960)
. Rayleigh scattering is interpolated from calculations of Coulson (1959)
. Absorption and scattering by aerosols are not included. The radiatively active low, middle, and high clouds (see Cloud Formation) are assigned different absorptivities and reflectivities per unit thickness. Cloud albedo and absorptivity are functions of cloud thickness, height, solid and liquid water content, water vapor content, and solar zenith angle, following Rodgers (1967)
. Multiple scattering effects of clouds are treated as in Lacis and Hansen (1974)
- The parameterization of longwave radiation follows the approach of Harshvardhan et al. (1987) . For absorption calculations, the broadband transmission method of Chou (1984)  is used for water vapor, that of Chou and Peng (1983)  for carbon dioxide, and that of Rodgers (1968)  for ozone. Longwave absorption is calculated in five spectral bands (with wavenumbers between 0 to 3 x 10^5 m-1), with continuum absorption by water vapor following Roberts et al. (1976) . Cloud longwave emissivities are treated as in Harshvardhan et al. (1987) . Clouds are assumed to be fully overlapped in the vertical (radiatively active clouds fill the grid box). See also Cloud Formation.
- Simulation of cumulus convection (with momentum transport) is based on the scheme of Arakawa and Schubert (1974)
, as implemented by Lord (1978)
, Lord and Arakawa (1980)
, and Lord et al. (1982)
. Mass fluxes are predicted from mutually interacting cumulus subensembles which have different entrainment rates and levels of neutral buoyancy that define the tops of the clouds and their associated convective updrafts. In turn, these mass fluxes feed back on the large-scale fields of temperature, moisture, and momentum (through cumulus friction). The effects of phase changes from water to ice on convective cloud buoyancy are also accounted for, but those of convective-scale downdrafts are not explicitly simulated.
- The mass flux for each cumulus subensemble is predicted from an integral equation that includes a positive-definite work function (defined by the tendency of cumulus kinetic energy for the subensemble) and a negative-definite kernel which expresses the effects of other subensembles on this work function. The predicted mass fluxes are positive-definite optimal solutions of this integral equation under a quasi-equilibrium constraint (cf. Lord et al. 1982
- A moist convective adjustment process simulates midlevel convection that originates above the PBL. When the lapse rate exceeds moist adiabatic from one layer to the next and saturation occurs in both layers, mass is mixed such that either the lapse rate is restored to moist adiabatic or saturation is eliminated. Any resulting supersaturation is removed by formation of large-scale precipitation (see Precipitation). In addition, if the lapse rate becomes dry convectively unstable anywhere within the model atmosphere, moisture and enthalpy are redistributed vertically, effectively deepening the PBL (see Planetary Boundary Layer). Cf. Suarez et al. (1983)  for further details.
- Four types of cloud are simulated: penetrative cumulus, midlevel convective, PBL stratus, and large-scale condensation cloud. Of these, the PBL, large-scale, and cumulus clouds above 400 hPa interact radiatively (see Radiation); these are assumed to fill the grid box completely (cloud fraction of 1).
- Large-scale condensation cloud forms in layers that are saturated. The penetrative cumulus cloud and midlevel convective cloud are associated, respectively, with the cumulus convection and moist adjustment schemes (see Convection). PBL stratus cloud forms if the specific humidity at the PBL top is greater than saturation, and the cloud-top stability criterion of Randall (1980)  is met. The base of this cloud is determined as the level at which the specific humidity of the well-mixed PBL is equal to the saturation value (see Planetary Boundary Layer).
- The PBL is parameterized as a well-mixed layer (turbulent fluxes linear in the vertical) whose depth varies as a function of horizontal mass convergence, entrainment, and cumulus mass flux determined from the convective parameterization (see Convection). The PBL is identical to the lowest model layer, and its potential temperature, u-v winds, and specific humidity are prognostic variables. Turbulence kinetic energy (TKE) due to shear production is also determined by a closure condition involving dissipation, buoyant consumption, and the rate at which TKE is supplied to make newly entrained air turbulent. Because of the absence of vertical diffusion above the PBL (see Diffusion), discontinuities in atmospheric variables that may exist at the PBL top are determined from "jump" equations.
- The presence of PBL stratocumulus cloud affects the radiative parameterizations (see Cloud Formation and Radiation), the entrainment rate (through enhanced cloud-top radiative cooling and latent heating), and the exchange of mass with the layer above the PBL as a result of layer cloud instability. If the PBL lapse rate is dry convectively unstable, an adjustment process restores stability by redistributing moisture and enthalpy vertically (see Convection). Cf. Suarez et al. (1983)  and Randall et al. (1985)  for further details. See also Surface Characteristics and Surface Fluxes.
- The surface roughness lengths are specified as uniform values of 2.0 x 10^-4 m over ocean, 1 x 10^-4 m over sea ice, and 1 x 10^-2 m over continental ice. The roughness lengths over land vary monthly according to 12 vegetation types (cf. Dorman and Sellers 1989
), with daily values determined by linear interpolation.
- The snow-free land albedo varies monthly according to vegetation type, with daily values determined by linear interpolation. Albedos of ocean, ice, and snow-covered surfaces are prescribed and do not depend on solar zenith angle or spectral interval.
- Longwave emissivity is prescribed as unity (blackbody emission) for all surfaces.
- Surface solar absorption is determined from the albedos, and longwave emission from the Planck equation with prescribed emissivity of 1.0 (see Surface Characteristics).
- Turbulent eddy fluxes of momentum, heat, and moisture are parameterized as bulk formulae with drag/transfer coefficients that depend on vertical stability (bulk Richardson number) and the locally variable depth of the PBL normalized by the surface roughness length (see Surface Characteristics), following Deardorff (1972)
. The requisite surface atmospheric values of wind, dry static energy, and humidity are taken to be the bulk values of these variables predicted in the PBL (see Planetary Boundary Layer). The same exchange coefficient is used for the surface moisture flux as for the sensible heat flux.
- The surface moisture flux also depends on an evapotranspiration efficiency factor beta, which is set to unity over ocean and ice surfaces, but which is equal to a prescribed soil wetness fraction (see Land Surface Processes) that depends on vegetation type (cf. Dorman and Sellers 1989 ).
- Ground temperature is determined from a surface energy balance (see Surface Fluxes) without inclusion of soil heat storage (cf. Arakawa 1972
- Soil moisture (expressed as a wetness fraction) is prescribed monthly from climatological estimates of Mintz and Serafini (1981 ). Precipitation and snowmelt therefore do not influence soil moisture, and runoff is not accounted for; however, the prescribed soil moisture does affect surface evaporation (see Surface Fluxes).
Last update November 19, 1996. For further information, contact: Tom Phillips ( email@example.com)