Sensitivity of Arctic Sea Ice Thickness to Intermodel Variations in the Surface Energy Budget - Arctic Sea Ice Decline: Observations, Projections, Mechanisms, and Implications - page 74

Sensitivity of Arctic Sea Ice thickness to Intermodel Variations
in the Surface Energy Budget

Eric t. deWeaver

Center for Climatic Research, University of Wisconsin-Madison, Madison, Wisconsin, USA

Elizabeth c. Hunke

Los Alamos National Laboratory, Los Alamos, New Mexico, USA

Marika M. Holland

National Center for Atmospheric Research, Boulder, Colorado, USA

Sea ice simulations from an ensemble of climate models show large differences
in the mean thickness of perennial Arctic sea ice. to understand the large thickness
spread, we assess the sensitivity of thickness to the ensemble spread of the surface
energy budget. Intermodel thickness and energy flux variations are related through
a diagnostic calculation of thickness from surface temperature and energy fluxes.
the calculation shows that an ensemble range of 60 W m -2 in energy fluxes, as
simulated by climate models, results in an approximate range of 1-5 m in ice
thickness. The ensemble mean value of the melt season energy flux, together with
a budget residual term that represents the effects of ocean heat exchange and ice
divergence, are the key factors that determine the range in ice thickness owing to
the flux spread. The ensemble spread in summertime energy flux is strongly related
to the spread in surface albedo, while differences in longwave radiative forcing,
due in part to cloud simulation errors, play a smaller role.

one metric of this uncertainty is the range of simulated val-
ues for the mean thickness of present-day perennial Arctic
sea ice: 1 to 6 m (excluding one outlier) over the simula-
tion ensemble of “20th century climate in coupled models”
(20c3M) [e.g., Randall et al. , 2007]. Improved understand-
ing of the factors determining this range and the require-
ments for reducing it through model improvement would be
of some interest, particularly given the use of climate model
projections for Arctic-related policy decisions [e.g., Amstrup
et al. , this volume].
Simulation assessments prepared for policy makers [e.g.,
Kattsov et al. , 2005; Randall et al. , 2007; DeWeaver , 2007]
typically account for sea ice simulation uncertainty by de-
scribing the delicate physical processes which determine ice

1. IntroductIon

the recent dramatic decline in September Arctic sea ice
extent has led to increased interest in climate model projec-
tions of future Arctic sea ice conditions. However, future sea
ice projections are subject to substantial uncertainty, as can
be seen by comparing results from different climate models.