Sea Ice–Albedo Feedback and Nonlinear Arctic Climate Change - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 117

−9°C. It is noteworthy that this loss of ice does not alter the
nature of the decline in effective annual albedo in either
model. This behavior was also noted in the 21st century plot
(Plate 3). Here we see that the extension to warmer tempera-
tures does involve nonlinear effective albedo changes: steep-
ened arcing in the CCSM3 and a kink-like turn in ECHAM5.
The total fall in effective surface albedo is over 0.5 in both
models, but the effective planetary albedo drop over the ex-
periments is about 0.1 for both models (not shown), indi-
cating a large role for atmospheric shortwave masking and
shortwave property changes.
The rapidity of the transition to annually ice-free condi-
tions in the ECHAM5 model and the failure of subsequent
variability to produce significant ice are suggestive of an un-
stable transition to a new equilibrium. Since the lifetime of
sea ice in the Arctic (about 10 years) is short compared to the
timescale of CO 2 increase (70 years for CO 2 doubling), we
can view the ice as passing through a series of quasi-equili-
brated states as the warming progresses. Under this interpre-

tation the rapid transition to the annually ice-free state in the
MPI model bears some resemblance to the SICI of simple
energy balance models which occurs abruptly as a global
forcing is gradually raised above a threshold value.
To explore further the connection between the transition
and SICI, we look at the changes in surface albedo feed-
back (SAF) as the transition progresses. using the fact that
the model transitions are more similar in temperature than
in time (Plate 4), we evaluate the surface albedo feedback
in three (annual mean) temperature eras: −15°C to −10°C
(perennial to seasonal ice transition), −10°C to −5°C (sea-
sonal ice), and −5°C to 0°C (transition to ice free). A method
of Winton [2005] is used to estimate the SAF. This method
uses standard model output to fit a simple optical model and
estimate the impact of surface albedo changes on shortwave
absorption.
In both models, surface albedo feedback makes an increas-
ing contribution to the decline in sea ice as air temperatures
approach freezing (Figure 1, top). In the NCAR model the

Figure 2. Polar atmosphere heat balance changes over three temperature eras: (top left) top-of-atmosphere absorbed
shortwave and (top right) outgoing longwave radiation and (bottom left) atmospheric heating from sides and (bottom
right) upward heat flux from the surface.