Arctic Cloud Properties and Radiative Forcing From Observations and Their Role in Sea Ice Decline Predicted by the NCAR CCSM3 Model During the 21st Century - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 56

decades. In CCSM3, increased cloud cooling of the surface
in the end of the 21st century only partly compensates for
the large surface albedo decrease. The largest increase in the
net SW flux by 24 W m -2 is found in June. At the same time,
a large increase in the LW cloud forcing is simulated during
the entire year with a maximum of 52 W m -2 in December
and minimum of 10 W m -2 in July (Figure 11a).
Cloud changes are often described in terms of cloud
fraction. The CCSM3 model indeed predicts an increase
in both cloud fraction and cloud LWP during the 21st cen-
tury. However, magnitudes of the increase differ depending
on the month, and in some months comparable increase in
cloud fraction can be accompanied by different increase in
cloud LWP and as a consequence different changes in down-
welling LW and SW fluxes (Table 1). Changes in the net SW
flux, shown in Table 1, also strongly depend on the insola-
tion and surface albedo.

even a small increase in the cloud liquid water content sig-
nificantly increases the downwelling longwave flux. Cloudy
skies in winter are always associated with warmer surface
temperatures because of the efficient energy transfer from
the relatively warm cloudy atmosphere to the surface by
increasing the downwelling LW flux. CCSM3 simulations
for the 21st century predict an increase in the cloud liquid
water content together with warming of the atmospheric
boundary layer. The largest increase in the cloud liquid
water content and near-surface atmospheric temperature
is predicted during the winter with a large impact on the
downwelling longwave flux. Thermodynamic model studies
showed that this is the time when cloud changes have the
largest impact on the sea ice thickness [ Curry et al. , 1993;
Shine and Crane , 1984].
In summer, clouds are present practically continuously
and contain large amounts of liquid. Above certain threshold,
clouds emit as blackbodies, and further increase in the cloud
liquid content has no effect on the cloud longwave forcing.
For this reason, increase in the cloud base temperature plays
a more important role for downwelling longwave flux during
the summer. Cloud cooling effect also becomes significant
during the summer because of both high cloud optical thick-
ness and thus high cloud albedo and the lowered surface
albedo over melting sea ice. Clouds reduce monthly mean
shortwave radiation reaching the surface by up to 100 W m -2
on average over the Arctic Ocean during the summer. At the
same time, an increase in the shortwave radiation absorbed
by the surface-atmosphere column up to the same magnitude
can occur when sea ice gives way to the open ocean for all-
sky conditions. Simulations with the CCSM3 model showed
that during the 21st century, clouds significantly diminish
but do not cancel the effect of reduced surface albedo on the
surface-absorbed shortwave flux.
The relative role of the shortwave and longwave cloud
forcing during the 21st century depends strongly on how
the model simulates cloud properties. Mixed phase cloud
parameterization in the CCSM3 model allows liquid water
to be present at temperatures between -10 and -40°C. To-
gether with excessive poleward moisture flux, this leads to
very high liquid water content in the Arctic clouds, which
is overestimated compared to the SHEBA ground-based
observations. There are not enough Arctic-wide observa-
tions to say if this overestimation is significant. Part of the
model's overestimation of the cloud LWP comes from the
fact that CCSM3 values are averaged over the ocean and sea
ice areas north of 70°N, while SHEBA measurements are
from perennial ice area. In order to improve model simula-
tion of cloud properties, there is a need for more year-round
ground-based observations of clouds and radiative fluxes for
various sea ice conditions. Correctly predicting seasonality

5. SUMMARY AND DISCUSSION

Drastic sea ice retreat has been observed in the Arctic dur-
ing the last decade of the 20th century and beginning of the
21st century. Most global climate models forced with to-
day's trends in atmospheric greenhouse gas concentrations
predict drastic sea ice decline in the Arctic by the end of the
21st century. The response of the Arctic climate system to
initial warming is not linear and involves multiple feedbacks
able to accelerate or diminish the surface warming. The sea
ice-albedo feedback is considered one of the main factors
accelerating sea ice disappearance by increasing the amount
of absorbed solar radiation at the surface [see Winton , this
volume]. Increased cloud formation is thought to mitigate
the Arctic warming by replacing the highly reflective sea
ice surface. However, clouds also contribute to the surface
warming by increasing the downwelling longwave flux,
thus enhancing the greenhouse effect. In the present chap-
ter, we examined satellite and ground-based observations in
attempt to disentangle cloud effects on the shortwave and
longwave radiative fluxes. We also discussed changes in sea
ice, clouds and surface radiative fluxes predicted by the cou-
pled global climate model NCAR CCSM3 during the 21st
century.
Over the Arctic perennial sea ice, clouds have a net warm-
ing effect on the surface during most of the year by increas-
ing the downwelling longwave flux. The magnitude of the
downwelling longwave flux strongly depends on cloud
properties, such as cloud liquid water content and cloud base
temperature. Clouds in the Arctic are usually mixed phase
with frequent presence of liquid during winter and continu-
ous large amounts of liquid during the summer. During win-
ter, when cloud events are occasional and clouds are thin,