1. The atmospheric model The basic features of the atmospheric and land surface parts of GCMII are similar to those of the earlier version (GCMI, described in some detail in Boer et al., 1984). In particular, the spectral formulation for representation of the horizontal variation of prognostic variables is retained. Important differences between the models are briefly described in the following subsections. Vertical discretization The vertical discretization of the prognostic equations in GCMII differs from that in GCMI. A hybrid vertical coordinate and a finite-element formulation (both discussed in Laprise and Girard, 1990) are used. This formulation has a number of advantages over the vertical finite-difference scheme used in GCMI, including flexibility in the choice of different layering schemes for thermodynamic and momentum variables and the conservation of energy and angular momentum in the absence of physical sources or sinks. GCMII has ten vertical levels and employs a triangular spectral truncation having 32 longitudinal waves (T32/L10). Moisture variable Specific humidity is the prognostic moisture variable in GCMII, while dewpoint depression is used in GCM1. This change of moisture variable was motivated mainly by the desire to ensure that the discretized prognostic equation for moisture is conservative in the absence of sources and sinks of water vapour. Such a conservation principle cannot be assured when dewpoint depression is used as the moisture variable. Parameterization of unresolved transfer processes, precipitation, and latent heat release Representation of the effects of unresolved transfer processes and the generation of precipitation and latent heat release in GCMII are similar in many respects to those used in GCM1 (as outlined in sections 5 through 7 of Boer et al., 1984). The moist convective adjustment and large-scale precipitation algorithms are the same as those employed in GCMI. However, the parameterization of vertical transfer processes at the surface and in the free atmosphere have been modified to some extent. Turbulent vertical fluxes at the surface and in the free atmosphere Vertical fluxes of momentum, heat and moisture due to turbulent transfer processes are represented using eddy diffusivity formulations in the free atmosphere, while those at the surface are represented in terms of drag coefficient formulations. These formulations are basically the same as those used in GCM1, the main differences being in the functional dependence on stability parameters (gradient and bulk Richardson numbers) and the value of the outer mixing-length scale used in the eddy diffusivity formulation. Estimation of the temperature at screen level To depict the effect of warming due to increased amounts of CO2, it is common to use either the temperature at the lowest model level or the surface temperature. These two temperature are usually different from each other, and neither is consistently more representative of the air temperature near the surface. The observed variable is, of course, the temperature at the screen level (2 m above the surface). The version of GCMII used for control and doubled CO2 experiments has ten levels in the vertcal with the lowest prognostic level located at approximately 200 m above the surface. GCMII uses a gradient profile relationship to estimate the air temperature at the screen level. The required gradient profile is obtained by noting that, in accordance with surface-layer theory, the vertical heat flux is not a function of height in the region between the surface and the screen level. In this region the diffusive representation used in the free atmosphere and the bulk formulation used at the surface are consistent with each other. Orographic gravity-wave drag The orographic gravity-wave drag parameterization used in GCMII is described in detail in McFarlane (1987). In particular, the effects of breaking and dissipation of unresolved orographically excited gravity waves is represented as an additional drag force on the resolved flow. Surface energy balance and hydrology over land The treatment of surface processes over land has been modified extensively in GCMII. A single soil-layer is used as in GCMI, but the properties of this layer now vary with location. In order to obtain more realistic simulations of the diurnal variation of surface temperatures the energy storage in the soil is represented using the force-restore method rather than the thermal intertia method used in GCMI. As in GCMI the soil moisture regime is represented in terms of a simple "bucket" model. However, an important difference is that, in GCMII, the field capacity varies with location. Budget equations for both liquid and forzen forms of soil moisture are carried out in the model, as dicussed in section 8b of Boer et al., 1984. The land surface scheme in GCMII does not explicitly model the vegetation canopy. However, some of the effects of a vegetative canopy are represented in an approximate way by assigning spatially variable soil depths and evapotranspiration slope factors, with values being specified for each vegetation class. Clouds and radiation Cloud coverage In GCMII an interactive cloud scheme replaces the prescribed clouds of GCMI. The optical properties of the clouds are also interactive variables. The fractional cloud cover is evaluated from the prognostic moisture and temperature fields through relative humidity. Cloud optical properties The optical properties of clouds are evaluated from the cloud liquid water content (LWC). In the current version the LWC is specified to be propotional to the adiabatic liquid water content that results when saturated air at ambient temperature is lifted vertically through a small distance. Cloud albedo is calculated using the delta-Eddington method. Radiation The treatment of radiation in GCMII is different from that in GCMI. The new terrestrial radiation scheme follows the method developed by Morcrette (1984). The solar radiation scheme is an updated version of that described by Fouqart and Bonnel (1980) and used in GCMI. 1. Solar radiation The upward and downward solar irradiance profiles are evaluated in two stages. First the model calculates a mean photon optical path in a scattering atmosphere, including actual clouds, aerosols and Rayleigh diffusion. The reflectance and transmittance in clouds and aerosol layer are calculated using a delta-Eddington method (Joseph et al., 1976) and a two-stream approximation. Second, the scheme calculates the final downward and upward fluxes. 2. Terrestrial radiation Emission and absorbtion of terrestrial radiation is computed using a scheme originally developed by Morcrette (1984) and currently used in the ECMWF model (Morcrette 1990, 1991). An innovative feature of this scheme is a correction method that allows for adequate treatment of the pressure and temperature dependencies of the longwave line absorption. Surface albedo The mean surface albedo is specified for the two spectral intervals used in the solar radiation scheme. Over bare, dry land a local value is specified in each grid square as a weighted average of the values for each of the 23 vegetation categories of the Wilson and Henderson-Sellers (1985) data. These values are reduced by as much as 7% for wet soil. It is assumed that the land surface is covered with snow when sufficient snow mass has accumulated to give an average snow depth in excess of the snow masking depth. Over oceans each of the spectral intervals has the same albedo. This is specified as a function of latitude and varies monotonically from 6% in the tropics (between 30oN and 30oS)to 17% poleward of 70o latitude in both hemispheres. The ocean surface is taken to be ice covered when a sufficiently large mass of sea ice has accumulated. References: McFarlane, N.A., G.J. Boer, J.-P. Blanchet, and M. Lazare (1992): The Canadian Climate Centre Second-Generation General Circulation Model and Its Equilibrium Climate. Journal of Climate, Vol. 5, No. 10, 1013-1044. Boer, G.J., N.A. McFarlane, and M. Lazare (1992): Greenhouse Gas-induced Climate Change Simulated with the CCC Second-Generation General Circulation Model. Journal of Climate, Vol. 5, No. 10, 1045-1077. Boer, G.J., N.A. McFarlane, R. Laprise, J.D. Henderson, and J.-P. Blanchet (1984): The Canadian Climate Centre Spectral atmospheric general circulation model. Atmos. Ocean, 22, 397-429. Laprise, R., and C. Girard (1990): A spectral general circulation model using a piecewise-constant finite-element representation on a hybrid vertical coordinate system. J. Climate, 3, 32-52. McFarlane, N.A. (1987): The effect of orograpically excited gravity-wave drag on the circulation of the lower stratosphere and troposphere. J. Atmos. Sci., 44, 1175-1800. Morcrette, J.-J. (1984): Sur la parameterisation du rayonnement dans les modeles de la circulation generale atmospherique. Ph.D. thesis, l'Universite des sciences et techniques de Lille, 373pp. Fouquart, Y., and B. Bonnel (1980): Computation of solar heating of the Earth's atmosphere: A new parameterization. Beitr. Phys., 53, 35-62 Joseph, J.H., W.J. Wiscombe, and J.A. Weinman (1976): Thde delta-Eddington approximation for radiative flux transfer. J. Atmos. Sci., 33, 2452-2459. Morcrette, J.-J. (1990): Impact of changes to the radiation transfer parameterization plus cloud optical properties in the ECMWF model. Mon. Wea. Rev., 118, 847-873. Morcrette, J.-J. (1991): Radiation and cloud radiative properties in the ECMWF operational weather forecast model.Mon. Wea. Rev., 96, 9121-9132. Wilson, M.F., and A. Henderson-Sellers (1985): A global archive of land cover and soils data for use in general circulation models. J. Climatol., 5, 119-143. Hansen, J., A. Lacis, D. Rind, and G. Russell (1984): Climate sensitivity: Analysis of feedback mechanisms. Climate Processes and Climate Sensitivity, Geophy. Monogr., No. 29, AGU, 130-163.