Sensitivity of Arctic Sea Ice Thickness to Intermodel Variations in the Surface Energy Budget - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 84

Table 4. correlations Across the 20c3M Ensemble Between Sur-
face fluxes (radiative, Sensible, and latent), Albedo, and net Sur-
face radiation for the MJJA and JJA Seasons a
F SW ¯ F LW ¯

F SW F LW a (%) T M

MJJA

S 35 43 95 8 -91 72

a -9 -48 -90 -7 100 -77

F RAD 30 56 93 27 -90 75

Ŝ 14 30 59 19 -60 43

T M -24 82 61 43 -77 100

JJA

S 48 -26 94 -44 -85 74

a -9 19 -87 42 100 -64

F RAD 51 -17 96 -28 -85 67

Ŝ 28 -4 52 2 -45 32
T M 35 -21 68 -33 -64 100
a correlations are multiplied by 100. Values in F LW and F SW columns
are net fluxes.

is a prominent contributor to the ensemble spread in S , Ŝ ,
and hence h D , while longwave cloud effects play a smaller
role. Similar results are obtained if the correlations are taken
for the JJA months, excluding May from the melt season, as
shown in table 4 (bottom).

5. dIScuSSIon And concluSIonS

the conclusion of the diagnostic analysis is that the en-
semble spread in Arctic sea ice thickness is roughly con-
sistent with the spread in the surface energy budget, with a
range of 20 to 80 W m -2 in Ŝ (the melt season surface energy
fluxes plus a budget residual term) corresponding to a thick-
ness range of 1 to 6 m. the ensemble mean value of Ŝ is one
key factor in relating the ranges, and this value conforms
to observations to the extent that such comparisons can be
made. Another key factor is the difference in the lengths of
the ice growth and melt seasons. EUW's finding of incon-
sistency between the thickness and flux ranges stems from
their assumption of equal season lengths, rather than a short
(MJJA) melt season and a long growth season.

Figure 6. (a) Scatter of annual mean net longwave (positive up-
ward) and net shortwave fluxes (positive downward) for the 17
20c3M models, labeled by the model numbers in figures 3a and
5, with a solid line to denote exact balance between longwave and
shortwave fluxes. (b) Monthly sensible heat flux for the models,
with thick lines for models 1 and 2. (c) difference of surface air
temperature and surface temperature for all models, with thick lines
for models 1 and 2.