AMIP I/AMIP II Model Differences: Model DNM A5421
(4x5 L21) 1997
AMIP II Model Designation
Most Similar AMIP I Model
AMIP I/AMIP II Model Differences
AMIP II Model
Designation
DNM A5421 (4x5 L21) 1997
Most Similar AMIP
I Model
DNM
A5407.V2 (4x5 L7) 1995
AMIP I/AMIP II Model
Differences
Vertical Domain
The AMIP II model domain is from the surface to 10 hPa, a higher top than
for the AMIP
I model. The lowest atmospheric level for the AMIP II model is
at 993 hPa for a surface pressure of 1000 hPa, a lower level than for the
AMIP
I model.
Vertical Resolution
There are 21 unevenly spaced sigma levels in the AMIP II model, 3 times
the vertical resolution of the AMIP
I model.
Diffusion
In the AMIP II model, horizontal diffusion is computed on constant pressure
surfaces, rather than on sigma surfaces as in the AMIP
I model.
Gravity Wave Drag
Gravity wave drag is not simulated in the AMIP
I model. In the AMIP II model, drag associated with orographic
gravity waves is simulated after the method of Palmer et al. (1986)[43],
as modified by Miller et al. (1989)[48],
using directional independent subgrid-scale orographic variances obtained
from 1x1-degree topographic height data of Gates and Nelson (1975)[16].
Surface stress due to gravity waves excited by stably stratified flow over
irregular terrain is calculated from linear theory and dimensional considerations.
Gravity wave stress is a function of the atmospheric density, the low-level
wind, and the Brunt-Vaisalla frequency. The vertical structure of the momentum
flux induced by gravity waves is calculated from a local wave Richardson
number, which describes the onset of turbulence due to convective instability
and the turbulent breakdown approaching a critical level.
Chemistry
The AMIP
I model does not include the effects of aerosols. In the AMIP II model
the concentration of aerosols follows the zonal-average climatology of
Barker and Li (1995)[29].
Radiation
The radiation scheme of the AMIP
I model is replaced by the following representation:
-
Shortwave scattering/absorption is parameterized by the delta-Eddington
approximation of Joseph et al.(1976)[49]
and Galin (1998)[46] applied in 18 spectral
intervals, as described by Briegleb (1992)[50].
Following Slingo (1989)[33], the shortwave
optical properties of clouds for the delta-Eddington method are specified
for 4 spectral ranges with boundaries at 0.25, 0.69, 1.19, 2.38, and 4.0
microns. Optical properties of aerosols are fixed according to Barker
and Li (1995)[29].
-
Longwave absorption by carbon dioxide, water vapor, methane, nitrous oxide,
ozone, with transmission functions by Chou et al.(1991a[53],
1991b[52], 1993[51]),
is calculated for ten spectral intervals with boundaries at 0.0, 3.40,
5.40, 6.20, 7.20, 8.00, 9.80, 11.00, 13.80, 19.00, and 30.00 m-1.
-
Cloud liquid water is diagnostically determined following Lemus et al.
(1997)[31], and the ratio of cloud liquid
water and ice following Matveev (1984)[32].
Large-scale clouds are assumed to be fully overlapped within a single layer
(0-400 hPa, 400-700 hPa and 700-1000 hPa), but clouds in different layers
are treated as randomly overlapped. Convective clouds are assumed to be
fully overlapped.
-
Each grid box is divided into 8 sectors having the following characteristics:
no clouds, only upper clouds; only middle clouds; only low clouds; upper
and middle clouds; upper and low clouds; middle and low clouds; and upper,
middle, and low clouds together. In each sector, clouds are treated
as uniform in the horizontal direction with a fractional area that is determined
from the assumptions about the cloud vertical overlap. Radiative
fluxes are calculated separately for each sector, and the overall flux
for each grid box is computed as a weighted sum of the fluxes in the 8
sectors. See Galin (1998)[46] for further
details.
Convection
In the AMIP II model the Betts-Miller[42]
convection scheme is used in place of another convective adjustment scheme
in the AMIP
I model.
-
If the temperature of the lowest atmospheric level exceeds 287K, convection
after Betts and Miller (1984)[42] is
simulated as a relaxed convective adjustment towards calculated temperature
and humidity reference profiles that are based on observations. The relaxation
time is 12 hours for precipitating deep convection and 2 hours for nonprecipitating
shallow convection, which are regarded as mutually exclusive processes.
-
The convection is treated as shallow if the cloud top (defined as the level
of non-buoyancy) is below about 800 hPa for a surface pressure of 1000
hPa, or if there is insufficient moisture for precipitation to form; otherwise,
deep convection is operative.
-
If the temperature of the lowest atmospheric level is below 287K, then
convective adjustment after Manabe and Strickler (1964)[10]
operates. Once all types of convection are realized, the associated convective
mixing by the u and v winds is calculated.
Cloud Formation
The AMIP II model differs from the AMIP
I model in the following ways:
-
Large-scale cloud amount C at a particular vertical level is diagnostically
determined from relative humidity r: C = max[(r - rcr)/(1
- rcr), 0], where critical relative humidity rcr is specified
as 0.77 for high clouds (above 400 hPa), 0.75 for middle clouds (400 -
700 hPa), and 0.87 for low clouds (below 700 hPa). For the lowest layer,
the cloud fraction cannot exceed rcr.
-
Convective cloud formation is parameterized according to the amount of
convective precipitation, following Slingo (1987)[54].
Precipitation
-
In the AMIP II model convective precipitation is defined by the difference
in specific humidity before and after the occurence of Betts-Miller convection.
In the AMIP
I model, precipitation is determined from the Kuo moistening parameter
b.
-
In the AMIP II model evaporation of large-scale precipitation in a layer
below cloud is proportional to the saturation deficit and rain intensity.
Evaporation of convective precipitation is not treated. The AMIP
I model does not treat evaporation of falling precipitation at all.
Snow Cover
The AMIP II model differs from the AMIP
I model in the following ways:
-
Snow density, heat capacity, and heat conductivity depend on snow depth
and are defined as in Volodin and Lykossov (1997)[45].
-
The fraction of land covered by snow is 0.9 if the water-equivalent snow
depth is greater than 4x10-3m, and decreases linearly below
0.9 proportional to decreasing snow depth. Sublimation of snow is calculated
assuming saturated surface air humidity over the snow-covered area. The
surface roughness does not depend on the presence of snow. In the AMIP
I model fractional coverage of a grid box with snow is not represented
(snow covers the entire grid box).
Surface Characteristics
The AMIP II models surface characteristics differ from the AMIP
I model in the following ways:
-
Distinguished surface types are open ocean, land, continental ice, and
sea ice; each grid box includes only one of these types. If the surface
type is land, then the grid box is subdivided into areas occupied by snow,
bare soil, vegetation type, and canopy-intercepted or inland water.
The 11 vegetation types are as described by Dorman and Sellers (1989)[55],
a reduction of the 72 types of Wilson and Henderson-Sellers (1985)[58].
For dry conditions in the root-zone, the fractional area of vegetation
decreases, and that of the bare soil increases.
-
Albedo is altered by snow cover as a linear function of the ratio of the
water-equivalent snow depth to a critical value (4x10-3m). Albedos
of snow in open areas range from 0.5 to 0.8 and, in forested areas from
0.5 to 0.7. In the AMIP
I model snow-covered areas have albedo values ranging from 0.2 and
0.6.
-
Longwave emissivity is prescribed as 1.0 for ocean, 0.99 for ice and snow,
and 0.96 for land without snow. For partially snow-covered areas, the emissivity
varies linearly from 0.96 to 0.99 as the snow-water equivalent depth varies
from 0 to 4x10-3 m. Longwave emissivity is 1.0 everywhere
in the AMIP
I model.
Surface Fluxes
The AMIP II model differs from the AMIP
I model in the following respects:
-
Surface turbulent eddy fluxes of momentum, heat and moisture are simulated
as stability-dependent bulk formulae, following Monin-Obukhov similarity
theory. For calculation of drag and exchange coefficients, empirical functions
of Businger et al. (1971)[19] are matched
with a -1/3 power-law dependence for strong instability (Kazakov and Lykossov,1982)[6];
these are used as universal functions of roughness length and thermal stability
to define the vertical profiles of wind, temperature, and humidity in the
constant-flux surface layer. The required near-surface values of wind,
temperature, and humidity are taken to be those at the lowest atmospheric
level.
-
The surface moisture flux depends on the surface specific humidity; over
ocean, snow, ice and wet vegetated fractions, this is taken as the saturated
humidity at the surface temperature and pressure. Over bare soil, the surface
specific humidity is the product of relative humidity and the saturated
specific humidity. Relative humidity is a function of soil moisture in
the upper 0.08 m of the soil column (see Land Surface Processes). For a
dry vegetation canopy, the potential evaporation is reduced by an efficiency
factor that is the inverse sum of aerodynamic and stomatal resistances.
The stomatal resistance depends on shortwave radiation, air temperature,
humidity, and soil moisture as in Sellers et al. (1986)[59]
and Dorman and Sellers (1989)[55]. See
also Surface Characteristics.
Land Surface Processes
The AMIP II model includes a more complex representation of the soil than
the AMIP
I model:
-
Parameterizations of soil processes follow the method of Volodin and Lykossov
(1997)[45]. Surface temperature is calculated
as a weighed sum of temperatures of 4 surface types: snow, bare soil, vegetation,
and canopy-intercepted or inland water. The temperature of each surface
type is calculated from a heat balance equation. In the soil column, equations
of heat and moisture diffusion are solved, including terms that treat mutual
heat/moisture diffusion and hydraulic water flux; melting/freezing of soil
water is also included. The diffusion equations are solved at 24 irregularly
spaced levels in the soil at depths from 0.01 m to 10 m, with zero-flux
conditions at the lower boundary. In the case of a snow pack, 3 additional
regularly spaced levels are used to calculate heat diffusion. Snow heat
conductivity depends on its depth, following Palagin (1981)[64].
-
Soil heat capacity depends on water and ice amount, and soil heat conductivity
on soil moisture (cf. McCumber and Pielke, 1981[63]).
Soil moisture conductivity and hydraulic flux depend on soil type as in
Clapp and Hornberger (1978)[60]. The
coefficient of mutual heat-moisture conductivity is defined as in Palagin
(1981)[64]. All coefficients vary
according to the 11 soil types (cf. Cosby et al., 1984)[61]
that are specified from data of Zobler (1986)[65].
-
Surface runoff depends on rain or snowmelt intensity, upper soil moisture
and ice content. Drainage from each layer depends on its moisture
capacity (cf. Dümenil and Todini (1992)[62]).
Absorption of water by roots also is treated. See also Surface
Characteristics and Surface Fluxes.
Go to:
References
Return to:
Top of Page
AMIP II Model/Experiment Documentation
AMIP II Documentation Directory
AMIP Home Page
Last update August 10, 1999. For questions or comments, contact
Tom Phillips (phillips@pcmdi.llnl.gov).
LLNL Disclaimers
UCRL-MI-135872