**PMIP Documentation for CCSR1**

**Center for Climate System Research:
Model CCSR/NIES 5.4 02 T21/L20 1995**

WWW URL: http://climate3.ccsr.u-tokyo.ac.jp/home.html
(in Japanese);

Number of days in each month: 30 30 30 30 30 30 30 30 30 30 30 30

The model is identical to the latest AMIP model exept for different
initial conditions and the Earth's orbital parameters.

A summary of model features including fundamental equations is provided
by Numaguti et al. (1995). The spectral formulation of atmospheric dynamics
follows closely Bourke (1988). The radiation scheme is described by Nakajima
and Tanaka (1986) and Nakajima et al. (1996). The convective parameterization
is based on the work of Arakawa and Schubert (1974) and Moorthi and Suarez
(1992). Cloud formation is treated prognostically after the method of Le
Treut and Li (1991). Gravity-wave drag is parameterized as in McFarlane
(1987). The planetary boundary layer (PBL) is simulated by the turbulence
closure scheme of Mellor and Yamada (1974, 1982)[10,11]. The representation
of surface fluxes follows the approach of Louis (1979), with inclusio of
adjustments recommended by Miller et al. (1992) for low winds over the
oceans.

The radiative flux in each wavelength region is calculated as a sum of the products of the fluxes over the subchannels and their respective k-distribution weights, where each subchannel's fluxes are calculated by the two-stream DOM. The optical depth of each subchannel is estimated as the sum of the optical thicknesses of band absorption and of continuum absorption by gases. The transmissivity, reflectivity, and source function in each layer then are calculated as functions of optical depth, single-scattering albedo, asymmetry factor, cutoff factor, Planck function, solar incidence, and solar zenith angle. At each layer interface, fluxes are computed by the adding technique.

In the presence of clouds, radiative fluxes are weighted according to
the convective and large-scale cloud fractions of each grid box. The fluxes
are computed by treating clouds as a mixture of scattering and absorbing
water and ice particles in the shortwave; shortwave optical properties
and longwave in the emissivity are functions of optical depth. Fluxes therefore
depend on the prognostic liquid water content (LWC) as well as on the fraction
of ice cloud (see Cloud Formation). Radiative transfer in large-scale and
convective clouds is treated separately, assuming random and full overlap,
respectively, in the vertical. Cf. Numaguti et al. (1995) for further details.

The predicted convective mass fluxes are used to solve budget equations that determine the impact of convection on the grid-scale fields of temperature (through latent heating and compensating subsidence) and moisture (through precipitation and detrainment). The vertical mass flux at the base of each cloud type is predicted from the cloud work function A, defined as the integral over the cloud depth of the product of the mass flux (with a linear vertical profile assumed) and the buoyancy (proportional to the difference between the cloud virtual temperature and that of the grid-scale environment at the same height). Because a nonzero cloud-base mass flux implies a positive-definite work function, the former is determined assuming that the work function vanishes in a specified time scale T > that is longer than the convective time step t of 80 minutes.

In the RAS scheme, the new cloud-base mass flux at each time step is
estimated by the method of virtual displacement: the amount of grid-scale
warming and drying expected from a unit mass flux is calculated, and a
new cloud work function A' is determined; the new cloud-base mass flux
M' then is derived from a simple proportionality relation. The grid-scale
mass flux is obtained by summing over the contributions from the spectrum
of cloud types. The profile of the mass flux associated with convective
downdrafts is also simulated from a fixed fraction of the evaporation of
convective precipitation (see Precipitation). Updated values of convective
cloud fraction and convective liquid water content (LWC) for the grid box
also are determined from the grid-scale mass flux (see Cloud Formation
and Radiation). Cf. Numaguti et al. (1995) for further details.

The fraction of stratiform cloud C in any layer is determined from the
probability that the total cloud water (liquid plus vapor) is above the
saturated value, where a uniform probability distribution with prescribed
standard deviation is assumed. (For purposes of the radiation calculations,
the square root of C is taken as the cloud fraction). At each time step,
new values of LWC and vapor are determined by iteration, subject to conservation
of moist internal energy. The portion of C that is ice cloud is assumed
to vary as a linear function of the temperature departure below the freezing
point 273.15 K, with all of C being ice cloud if the temperature is <
258.18 K. Cf. Numaguti et al. (1995) and Le Treut and Li (1991) for further
details.

For 21fix ? ? ? ?

For 21fix run, sea ice edge of CLIMAP data is used for Feb. and Aug.,
the shape of sea ice edge for other months is determined by considering
summer SST over both hemisphere. Grid with lower summer SST is freeze earlier
and melt later.

Over ice surfaces, the albedo is a constant 0.7 (unaffected by snow
accumulation). Over ocean, the albedo depends on sun angle and the optical
thickness of the atmosphere. The albedos of the land surface are specified
according to vegetation type from the data of Matthews (1983). For snow-covered
land, the albedo increases over that of the background proportional to
the square root of the fractional snow cover. Longwave emissivity is everywhere
specified to be 1.0 (i.e., blackbody emission). See also Surface Fluxes
and Land Surface Processes.

The associated drag/transfer coefficients are functions of the surface roughness (see Surface Characteristics) and vertical stability expressed as a function of a modified Richardson number (see Planetary Boundary Layer). The effect of free convective motion is incorporated into the surface wind speed following Miller et al. (1992), and the surface wind speed also is not allowed to fall below 4 m/s.

For calculation of the moisture flux over ocean, ice, and snow-covered
surfaces, the evaporation efficiency beta is unity. Over partially snow-covered
grid boxes, beta increases as the square root of the snow fraction (see
Snow Cover). The evaporation efficiency over vegetation is limited by the
specified stomatal resistance. Cf. Numaguti et al. (1995) for further details.
See also Land Surface Processes).

Soil liquid moisture is predicted in a single layer according to the "bucket" formulation of Manabe et al. (1965)) The moisture field capacity is a spatially uniform 0.15 m, with surface runoff occurring if the predicted soil moisture exceeds this value. Snowmelt contributes to soil moisture, but if snow covers a grid box completely, the permeability of the soil to falling liquid precipitation becomes zero. For partial snow cover, the permeability decreases proportional to increasing snow fraction (see Snow Cover).

Soil moisture is depleted by surface evaporation; the evaporation efficiency beta (see Surface Fluxes) is not determined solely by the ratio of soil moisture to its saturation value, but is limited by the specified stomatal resistance of the vegetation. Other effects of vegetation, such as the interception of precipitation by the canopy and its subsequent reevaporation, are not included. See also Surface Characteristics and Surface Fluxes.

Last update November 9, 1998. For further information, contact: Céline Bonfils (pmipweb@lsce.ipsl.fr )