The Atmospheric Response to Realistic Reduced Summer Arctic Sea Ice Anomalies - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 94

is given by Alexander et al. [2004]. The modeling results
are generally presented as monthly anomalies constructed
by averaging over the 51 ensembles and subtracting the cor-
responding long-term monthly mean over the last 51 years
of the control simulation.
The discussion in this paper focuses on the model response
during August of 1995. Ice anomalies evolve during the sim-
ulation based on observed April to October ice conditions. In
1995, ice was below normal in the Kara-Barents seas in July
and throughout the Eurasian Arctic and in the Chukchi and
Beaufort seas during August (Plate 1) and September. The
atmospheric response in June and July was generally weak
and will not be discussed. This may be a consequence of
overall smaller sea ice anomalies during these months.

[ Briegleb and Bromwich , 1998a, Figure 10b], but the cloud
water path is too high resulting in clouds that reflect (emit)
excessively in the shortwave (longwave) range [also see
Gorodetskaya and Tremblay , this volume].
While the model has some deficiencies over the Arctic,
e.g., it is colder and wetter than observed (which also oc-
curs in most other atmospheric general circulation models
[ Randall et al. , 1998]), many aspects of the Earth's climate
are well simulated. This is a well-documented model that
has been used in numerous studies of the impact of sea ice
on the atmosphere [e.g., Deser et al. , 2004; Magnusdottir et
al. , 2004; Alexander et al. , 2004].

2.4. Linear Baroclinic Model

2.3. Atmospheric General Circulation Model To understand the mechanism for the large-scale response

over the North Pacific to reduced Arctic sea ice in August, we
forced a LBM with daily mean diabatic heating and transient
eddy heat and vorticity fluxes, similar to Deser et al. [2007],
from Cntle and Sum95e. The LBM [see Peng et al. , 2003] is
based on the primitive equations configured with T21 hori-
zontal resolution and 10 equally spaced pressure levels from
950 to 50 hPa. The model is linearized about the CCm3 ba-
sic state obtained from the long-term August mean in Cntle.
The LBM includes dissipation in the form of Rayleigh fric-
tion in the momentum equation and Newtonian cooling in
the thermodynamic equation, as well as biharmonic thermal
diffusion. The Rayleigh and Newtonian damping time scales
are 1 day at 950 hPa transitioning linearly to 7 days above
700 hPa. The LBM is integrated for 31 days.
The pattern of the CCM3 response to sea ice forcing is
diagnosed by comparing the LBM responses to anomalous
diabatic heating and transient eddy fluxes from Cntle and
Sum95e. The transient eddies are based on 14-day high-
pass-filtered data, constructed by subtracting the 11-day run-
ning means from the raw daily data (the half-power point of
this filter is 14 days).

The CCM (version 3.6) is the atmospheric GCM used in
this study; it has 18 vertical levels and a horizontal spec-
tral resolution of T42, which is approximately 2.8° latitude
by 2.8° longitude. Kiehl et al. [1998] describe the model
physics, while Hack et al. [1998] and Hurrell et al. [1998]
evaluate the model's climate with a global perspective while
Briegleb and Bromwich [1998a, 1998b] evaluate the polar
climate. Model evaluations relevant for this study will be
briefly outlined.
Hurrell et al. [1998] find that while the subtropical sum-
mer time SLPs are higher than observed, CCM3 captures the
key interseasonal shifts of the subtropical highs. The newer
Community Atmosphere Model (CAM3) displays similar
SLP features as CCM3 [ Hurrell et al. , 2006]. SLP over the
Arctic is higher than observed, and none of the Atmospheric
Model Intercomparison Project models investigated by Bitz
et al. [2002] capture the observed closed low over the central
Arctic during summer. An investigation by DeWeaver and
Bitz [2006] shows that JJA Arctic SLP in the Community
Climate System Model, version 3 (CCSM3) is too high, a
feature that is particularly prominent at T42 resolution. They
find that in the model there is subsidence due to a thermally
direct mean meridional circulation while reanalysis data in-
dicate rising motion with an indirect Ferrel cell in the Arc-
tic. Consistent with these studies, the Cntle simulation SLP
is 5-7 hPa too high compared to the National Centers for
Environmental Prediction/National Center for Atmospheric
Research (NCAR) reanalysis over the Arctic during August
(not shown).
Briegleb and Bromwich [1998b] find that CCm3 summer
time tropospheric temperatures in the Arctic are cooler than
observed by 2°-4°K, while precipitation minus evaporation
(P − E) compares favorably with observations. The Arctic
July total cloud amount in CCM3 is similar to observations

3. RESULTS AND DISCUSSION

3.1. Local Arctic Response

The model atmosphere displays a local thermal response
to reduced western Arctic sea ice extent. The net heat flux
anomalies resulting from the reduced sea ice are 10-25 W m -2
from the ocean to the atmosphere (Plate 2a). The sensible
heat flux is the dominant form of heating contributing about
4-8 W m -2 , followed by latent heat flux at 2-6 W m -2 and
then longwave at 2-4 W m -2 . Increased upward (downward)
directed longwave radiation of 2 W m -2 is associated with a
decrease (enhanced) in low-level clouds of 2%. It is not sur-