The first generation atmospheric general circulation model (AGCM1) is no longer used at CCCma. The following details about the model are listed here for historical purposes and to help the reader understand how our various models have evolved from their predecessors.

1. Model features

The first generation atmospheric general circulation model evolved from an earlier 5-layer version discussed by Boer and McFarlane (1979). The spectral formulation in the CGCM makes use of a truncated expansion in spherical harmonics to represent model variables in the horizontal.

Other features of the numerics include semi-implicit time-stepping (Robert et al., 1972) with a weak time filter (Asselin, 1972). The basic structure of the model is similar to that of the spectral forecast model of Daley et al. (1976), although some improvements have been made in the procedure for implementing the spectral algorithms and, of course, important additional physical processes have been included.

The equation governing horizontal motion are written in terms of vorticity and divergence of the horizontal wind. The remaining basic prognostic equations include the themodynamic equation written in terms of a function of geopotential height, the moisture equation written in terms of dew-point depression, and the surface pressure equation. Temperature is determined diagnostically from the geopotential via the hydrostatic equation, and the vertical motion variable is determined from the mass continuity equatiovia the hydrostatic equation, and the vertical motion variable is determined from the mass continuity equation.

The physical processes that govern the hydrological cycle include prognostic equations for soil moisture and snow accumulation.

Horizontal fluxes due to unresolved motions are parameterized following the suggestion of Leith (1971) which is particularly suitable for traingularly truncated spectral models.

A novel feature of the model is the inclusion of a parameterization of the momentum transport and deposition (wave drag) induced by topographically generated vertically propagating gravity waves.

2. Boundary conditions

While initial conditions are unimportant for general circulation models, boundary conditions have an important effect on the simulated climate. The effect of large-scale topography on the smilated climate enters into the equations of motion through the specification of the geopotential height at the surface. The values used in the model are obtained by truncating the spherical harmonic series representation of the topographic height at triangular 20-wave truncation.

Surface temperatures are computed over land and oceanic pack-ice by solving the surface energy balance equation. Surface temperatures are specified over the oceans using monthly mean climatological sea surface temperatures provided by the Marine Environmental Data Services Branch of the Department of Fisheries and Oceans.

The parameterization of surface fluxes over land requires the specification of bulk transfer coefficients at neutral stability. The drag coefficient field is used for this purpose following Cressman (1960).

Surface albedo over land has a prognostic component that depends on snow cover. Background values of albedo over land, ocean and pack-ice obtained from Posey and Clapp (1964) are used.

3. Radiative transfer processes

Solar and terrestrial radiation provides the primary energy source and sink for the climate system. In the atmospheric general circulation model described here the specification of ocean surface temperatures implies that the fluxes of heat and moisture from the oceans do not depend directly on the radiative balance of the ocean surface. Thus the model is not sensitive to the radiative calculations as it would be if an interactive ocean were present.

The radiative processes that are included in the model result in heating or cooling in each atmospheric layer and at the surface which may be land or pack-ice. The radiative calculations are performed for two broad spectral regions - solar and thermal.

4. Horizontal transfer processes

Major horizontal transfers in the atmosphere are accomplished by the large-scale flow, which is explicitly calculated in the model. Nevertheless, the effect of unresolved horizontal scales of motion on those that are explicitly resolved in the model must be included in the formation if the results are to be realistic.

While a complex physical system like the atmosphere cannot be expected to display exactly a simple turbulent behaviour, the approximate correspondence suggests that, in the absence of complete knowledge and theory concerning atmospheric behaviour, the turbulences concepts may provide useful guidance in parameterizing the effects of subgrid-scale processes on scales explicitly resolved in the model.

This was the attitude adopted for AGCM1. In particular, the horizontal diffusion introduced in the model follows the work of Leith (1971), which provides a dissipative formulation that is obtained assuming the truncation wave number of the model falls within the enstrophy cascading inertial subrange.

5. Precipitation and latent heat release

In the model, precipitation occurs and latent heat is released when the local relative humidity becomes large enough so that supersaturation can occur in a given atmospheric column. The latent heat release may be associated with moist convection when the atmosphere is locally conditionally unstable. Both condensation and convection are treated by a convective adjustment scheme and applied to individual atmospheric columns. All condensed liquid water falls to the surface as precipitation.

6. Surface energy balance and hydrology

In the model the surface of the earth may be bare or snow-covered soil, glacial or sea ice, or open ocean. Surface temperatures are specified as a function of time over the open ocean. In all other cases surface temperatures are determined so as to satisfy the requirements of a surface heat balance.

The soil is considered to be completely snow covered when the snow mass per unit area exceeds a certain specified value. Surface albedo is allowed to depend on snow cover. Soil wetness and snow mass are prognostic variables in the model.

Over pack-ice, snow is evaporated first but afterward the pack itself acts as an infinite reservoir of frozen water for evaporation. Gloacial ice-packs such as those over Greenland and Antarctic subcontinent are represented as thick snow layers for surface hydrology calculations. Since the solar forcing for the simulation includes annual and diurnal variations, these snow layers never melt away.

Runoff is not explicitly calculated. When the total soil moisture reaches a value in excess of the field capacity the excess is assumed to run off and soil moisture is reset to unity. When rain falls on pack-ice it is assumed to run off immediately.

References:

Asselin, R., 1972: Frequency filter for time integrations. Mon. Weather Rev., 100, 487-490.

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(4), 397-429.

Boer, G. J., and N. A. McFarlane, 1979: The AES atmospheric general circulation model. Report of the JOC Study Conference on Climate Models: Performance, Intercomparison and Sensitivity Studies, Vol. I, GARP Publ. Ser. No. 22, pp. 409-460.

Cressman, G. P., 1960: Improved terrain effects in barotropic forecasts, Mon. Wea. Rev., 88, 327-342.

Daley, R., C. Girad, J. Henderson, and I. Simmonds, 1976: Short term forecasting with a muti-level spectral primitive equation model. Part I - Model formulation. Atmosphere, 14, 98-116.

Leith, C. E., 1971: Atmospheric predicability and two dimensional turbulence. J. Atmos. Sci., 28, 145-161.

Possey, J. W., and D. F. Clapp, 1964: Global distribution of normal surface albedo. Geofis. Int., 4, 33-48.

Robert, A. J., J. Henderson, and C. Turnbull, 1972: An implicit time integration scheme for baroclinic models of the atmosphere. Mon. Wea. Rev., 100, 329-335.