Some Aspects of Uncertainty in Predicting Sea Ice Thinning - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 63

Plate 1. Hence, if ice-albedo feedback were the cause, one
might anticipate that CCSM3 has above-average ice-albedo
feedback. This is advantageous because I intend to show that
ice-albedo feedback is too small to cause much uncertainty
in sea ice thickness change, even if it is uncertain within a
factor of 2.

2006a]. The polar climate of the 20th and 21st century in this
model is discussed by Meehl et al. [2006] and Holland et al.
[2006a]. The other components of CCSM3 and its climatol-
ogy are described in greater detail by Collins et al. [2006].

3.2. Experiments

3.1. Model Description To evaluate the effect of sea ice-albedo feedback on cli-

mate, pairs of experiments were run with the surface albedo
allowed to vary and with it held fixed over sea ice, and ocean
and the sensitivity to doubling CO 2 was evaluated. First, a
normal control integration was run with 1990s greenhouse
gas concentrations (CO 2 at 355 ppm). Then, a second con-
trol was run but with the surface albedo held fixed to a cli-
matological monthly mean annual cycle that was computed
from the normal control. This new “fixed-albedo control”
also had 1990s greenhouse gas concentrations. It was used
to verify that the method of fixing the albedo achieves a very
similar climate to the normal control. Finally, a pair of per-
turbed CO 2 experiments was run with CO 2 at 710 ppm: One
with the albedo free to vary and a second with the surface
albedo held fixed to the climatology from the normal con-
trol. Doubling CO 2 is representative of the anthropogenic
forcing level at about mid century in the Special Report on
Emissions Scenarios (SRES) A1B scenario. In all four runs,
the ocean heat transport is the same. Table 2 summarizes this
quartet of runs.
The surface albedo in CCSM3 is decomposed into two
spectral bands, denoted visible and infrared, for wavelengths
above and below 700 nm. These two bands are further de-
composed into direct and indirect beam components. For
simplicity, I will describe the method for fixing the albedo
as if there were only a single component, but in practice, this
method is applied to each of the four components separately.
Because the weighting of the four components depends on
clouds and atmospheric composition, which I am not fix-
ing, the total surface albedo in the grid cell may still vary
slightly.
The grid cell average albedo (for each albedo component)
is the weighted sum of the albedo for the sea ice-covered
fraction and the open water fraction:

All components of CCSM3 in the experiments are stand-
ard, except the ocean is a slab mixed layer rather than the full
ocean general circulation model, so the model can be run to
equilibrium in only a few decades. The slab ocean has depth
that is variable in space but fixed in time, and the ocean heat
transport is prescribed from a climatological monthly mean
annual cycle that was derived from a long control run of
CCSM3 with the standard ocean component.
The slab ocean and sea ice share a horizontal grid with
320 ´ 384 points, with resolution varying 0.5°-1° in the po-
lar regions. The Northern Hemisphere pole of this grid is
displaced to a point within Greenland. The atmosphere and
land model's horizontal resolution is truncated spectrally at
T42, and there are 26 vertical levels in the atmosphere.
The sea ice component in the experiments is distinct from
the very simple motionless sea ice component that is part of
the often used slab ocean option in the atmosphere compo-
nent of CCSM. Here, the sea ice model resolves a distribu-
tion of ice thicknesses using multiple ice categories, each
having a unique and variable concentration and thickness of
ice and snow and a unique surface energy balance, surface
albedo, and vertical temperature profile [ Bitz et al. , 2001;
Lipscomb , 2001]. The surface albedo is parameterized as a
function of snow depth, sea ice thickness, and surface tem-
perature. Melt ponds are not explicitly modeled, but their
influence is parameterized crudely by the dependence of the
albedo on temperature. The model momentum equation in-
cludes the elastic-viscous-plastic stress tensor of Hunke and
Dukowicz [2002]. The model also employs an explicit brine
pocket parameterization with shortwave radiative transfer
through the ice from Bitz and Lipscomb [1999]. The inclu-
sion of ice dynamics, multilayer thermodynamics, and an ice
thickness distribution have been shown to affect the climate
in coupled models [ Bitz et al. , 2001; Holland et al. , 2001, D D i A i D o 1 A i

(1)

Table 2. Experiments Conducted With CCSM3 in This Study
Experiment Name CO 2 , ppm Description

Normal control 355 control integration with freely varying albedo

Fixed-albedo control 355 sea ice and ocean albedo fixed to a climatology from the normal control

Normal perturbed 710 CO 2 raised to represent midcentury anthropogenic forcing

Fixed-albedo perturbed 710 CO 2 raised with sea ice and ocean albedo fixed to a climatology from the normal control