Multiple Equilibria and Abrupt Transitions in Arctic Summer Sea Ice Extent - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 167

Figure 17. (a) OHT time series used to illustrate differences under parameterizations of sections 3 and 4; (b) time series
of September ice area under section 3 parameterizations (solid curve) versus section 4 parameterizations (dashed curve);
(c) time series of March ice thickness, curves similarly labeled.

enon resembles the small ice cap instability found in other
simplified representations of sea ice described in section 2.
What is the source of these differences? The answer lies
in the parameterized ocean shortwave absorption. As long
as A n > 0, parameterizations (5) and (15) behave similarly
(solid and dashed curves in Figure 18) because of their
close relation through (17). In particular, the forms of the
two equilibrium solution branches connected by the sad-
dle-node bifurcation remain much the same. However, once
September ice cover is lost ( A n = 0), (5) implies a satura-
tion of shortwave absorption, whereas under (15) Q n can

continue to increase as ice cover in the other sunlit months
declines. The additional shortwave absorption becomes suf-
ficient to prevent freezing in winter, and Q n saturates at its
value for perennial absence of ice. In actuality, the impor-
tance of this effect is likely to be diminished by processes
not represented explicitly in our simple model, namely,
the strong sensible, latent, and longwave heat losses that
will occur from open water in winter. The loss of a much
larger fraction of the absorbed shortwave energy from the
sea surface under ice-free winter conditions than when
ice is present should tend to mitigate the large differences