Drivers of Arctic sea ice decline
in the HadGEM1 model
Ann Keen, Helene Hewitt and Jeff Ridley
FAMOS October 2013
© Crown copyright Met Office
Drivers of Arctic sea ice decline
in the HadGEM1 model
The factors driving the long term decline of the Arctic sea
ice are investigated using the heat budget of the ice and
overlying snow.
Talk outline:
• Motivation
• Brief description of the HadGEM1 model, scenarios used and
projections of future Arctic change
• Seasonal cycle of components of the heat budget of the Arctic ice and
overlying snow in the control integration
• Changes in the budget for the periods 2010-2019 and 2040-2049, and
significance compared to inter-annual variability
• Discussion of preliminary results, summary and future work
© Crown copyright Met Office
Motivation
Projections of September ice extent
from AR5 models (Stroeve et al, 2012)
© Crown copyright Met Office
HadGEM1: model description
• Global coupled GCM – results submitted to IPCC AR4 assessment
• Met Office UM atmospheric model, 1.25x1.875 degrees, 38 vertical levels.
• Ocean 1x1 degree resolution (increasing to 1/3x1 in the Tropics), 40 vertical
levels
• Sea ice component (1x1 degree resolution), largely based on the CICE model :
• Elastic-Viscous-Plastic (EVP) ice dynamics
• Ridging scheme to convert thinner ice to thicker ice and open water
• Sub-gridscale representation of the ice thickness distribution (ITD)
• Zero layer thermodynamics
• Albedo depends on surface temperature and snow cover
• Identified by a number of studies as having one of the better simulations of
present day Arctic sea ice compared to the other AR4 models.
• Used for a range of climate change experiments
© Crown copyright Met Office
HadGEM1: projections of
Arctic ice volume (September)
Control: fixed
pre-industrial
(1860) forcing
Scenario runs:
Time varying
anthropogenic
forcing:
observed to
2000, A1B or
A2 scenario
thereafter
© Crown copyright Met Office
Components of ice/snow heat
budget: HadGEM1 model (control)
Summer melting:
• Melting at the
top surface of the
ice during June
and July
• Basal melting
due to ocean to
ice heat flux,
extending into
Autumn
Melting
Freezing
Winter freezing:
• Diffusive heat flux through ice to the atmosphere
• Frazil ice formation
• ©Offset
byMetsome
basal melting
Crown copyright
Office
Changes in
components
of the heat
budget
2010-2019
(anthro+A1B w.r.t. control)
• Decline in ice volume
is due to extra melting
during the summer
• More surface melting
during June and July
• More melting at the
sides and base of the
ice during August
© Crown copyright Met Office
A robust
signal?
© Crown copyright Met Office
Looking further ahead
(2040-49)
• More surface
melting during
June, but not
July
2010-2019
response
© Crown copyright Met Office
• Even more
melting at the
sides and base
of the ice during
August and also
September
Interpretation:
surface melting changes
In HadGEM1 the ice albedo is
a function of surface
temperature.
Warmer surface temperatures
in June lead to reduced
albedo and extra melting
2010-2019
response
This effect is less evident in
July, as in the control
integration the surface
temperature is already close
to the melting temperature
Other changes (eg cloud)
under investigation
© Crown copyright Met Office
Interpretation: Ocean to ice
heat flux changes
It is likely that the extra
heating is due to in-situ
warming of the ocean surface
as the ice cover retreats:
• the extra melting is
confined to the summer
months
2010-2019
response
• spatially the increases in
o2i heat flux are
correlated with decreases
in ice concentration
Analysis of the heat budget of
the upper ocean is required to
confirm this.
© Crown copyright Met Office
Drivers of Arctic sea ice decline
in the HadGEM1 model:
Summary
• Changes in the sea ice heat budget in a number of climate change
integrations have been studied, to identify the factors driving the long
term decline in ice volume
• Most of the decline is due to extra ice loss during the summer melt
period
• There is extra surface melt in June, due to warmer surface
temperatures and an associated reduction in albedo
• As the ice retreats, in situ warming of the ocean surface results in an
increased ocean to ice heat flux in August, and eventually September
too.
• In both cases, the ice albedo feedback is important in determining
which months have extra melting – highlighting the importance of a
physically realistic representation of albedo in models used to make
projections of future climate change
© Crown copyright Met Office
Drivers of Arctic sea ice decline
in the HadGEM1 model:
Future work and opportunities
• Continuing analysis to confirm conclusions and quantify other
changes (eg role of clouds)
• Returning to the original motivation – this should be a useful approach
for inter-comparing the factors driving the decline in sea ice in a range
of models.
• Combined with other work to compare model budgets with
observations (West et al., in preparation), can help the assessment of
model strengths and weaknesses.
© Crown copyright Met Office
Questions and answers
© Crown copyright Met Office
HadGEM1: projections of
Arctic ice extent (September)
Control: fixed
pre-industrial
(1860) forcing
© Crown copyright Met Office
Scenario runs:
Time varying
anthropogenic
forcing:
observed to
2000, A1B or
A2 scenario
thereafter
Relationships between
ice extent and ice volume
• Thinner ice at start of
melt season will melt
away more readily over
the summer
• Thinner ice allows
greater loss of heat from
ocean to atmos: greater
autumn ice growth.
Area-weighted % of ice that melts away
completely over the summer, as a function
of April ice thickness. (HadGEM1 model)
© Crown copyright Met Office
• Thinner ice weaker
than thicker ice, and
more responsive to
dynamic forcing.
Title XXXXXXXX
Description of presentation
Speaker, location and date of presentation
© Crown copyright Met Office
Collaborators
© Crown copyright Met Office
Introduction
© Crown copyright Met Office
© Crown copyright Met Office