Title: A Simple Model for Arctic Sea Ice: Variability and Trends in September Minimum Ice Extent Kyle Armou
1A Simple Model for Arctic Sea Ice Variability
and Trends in September Minimum Ice ExtentKyle
ArmourUniversity of WashingtonDepartment of
PhysicsCecilia BitzUniversity of
WashingtonDepartment of Atmospheric
SciencesElizabeth HunkeLos Alamos National
Laboratory
2FY First-Year Ice MY Multi-Year Ice
W Winter maximum ice area S Summer minimum
ice area
All variables are for northern hemisphere
3FY First-Year Ice MY Multi-Year Ice
W Winter maximum ice area S Summer minimum
ice area f m f area of FY ice at summer
minimum m area of MY ice at summer minimum
4FY First-Year Ice MY Multi-Year Ice
- W Winter maximum ice area
- S Summer minimum ice area f m
- f area of FY ice at summer minimum
- m area of MY ice at summer minimum
- ? fraction of FY ice that survives summer melt
season - ? fraction of MY ice that survives summer melt
season
5FY First-Year Ice MY Multi-Year Ice
Simple model for summer minimum ice area in year
n
6FY First-Year Ice MY Multi-Year Ice
Simple model for summer minimum ice area in year
n
7mean state and trends (equilibrium and
slowly-varying component)
high-frequency variability (interannual
timescales)
8Equilibrium summer minimum ice area
If all parameters remain at their equilibrium
values, then
?
9(No Transcript)
10where are we??
11where are we??
12where are we??
13Equilibrium FY and MY summer minimum ice areas
summer minimum FY ice area
summer minimum MY ice area
14Equilibrium FY and MY summer minimum ice areas
summer minimum FY ice area
summer minimum MY ice area
15Equilibrium FY and MY summer minimum ice areas
summer minimum FY ice area
summer minimum MY ice area
16Sea Ice Simulation
Latest CICE ModelV4 from LANL (developed by Hunke
and Lipscomb) No Atmosphere - NCEP reanalysis
(hindcast) Mixed Layer Ocean - monthly
varying OHFC from CCSM run, annually
repeating First-Year and Multi-Year ice
categories FY ice area tracer (distinct from
the ice age tracer of Hunke and Bitz) All ice
that exists on September 15th each year is
promoted to the MY ice category Ice that
grows in open water after September 15th is
considered FY ice
171979-2006 March and September average ice extent
CICE hindcast - reproduces trend and
variability in September ice extent - has no
trend in March ice extent
Observations of ice extent (obs) from NSIDC
18March Exchange of FY and MY ice area
FY
area
MY
Anticorrelated FY has increased while MY has
decreased Little variability or trend in total
ice area
MY
R -0.7
106 km2
106 km2
FY
19September Independent FY and MY ice area
MY - m
area
FY - f
No correlation Only MY ice area has
declined Variability about the same, equally
important for variability in total ice area in
Summer
m
R -0.04
106 km2
106 km2
f
20September Independent FY and MY ice area
MY - m
area
FY - f
MY ice survival rate
?
?
FY ice survival rate
21FY and MY ice survival rates
MY - ?????????
FY - ?? 0.23
Not significantly correlated Trends in ? and ??
are comparable ( -0.02/decade)
?
R 0.3
?
22Larger trend in MY area than FY area
(f summer minimum FY ice area)
(m summer minimum MY ice area)
23Larger trend in MY area than FY area
MY - ?
MY - m
area
FY - ?
FY - f
24Sea Ice Age Model of Maslanik et al
25mean state and trends (equilibrium and
slowly-varying component)
high-frequency variability (interannual
timescales)
26Perturbation Analysis
27Perturbation Analysis
not significantly correlated winter ice area has
no significant influence on summer ice area
106 km2
R 0.10
106 km2
28Perturbation Analysis
not significantly correlated winter ice area has
no significant influence on summer ice area
106 km2
R 0.10
106 km2
29Perturbation Analysis
variance explained
6
106 km2
R 0.24
106 km2
30Perturbation Analysis
variance explained
6
69
37
106 km2
106 km2
R 0.83
R 0.61
31Perturbation Analysis
variance explained
69
37
6
- Summer minimum ice area depends almost entirely
on FY and MY ice survival rates (which are
functions of that years conditions - weather,
etc.) - Little year-to-year memory in summer
minimum area (similar to observations) - ?
is slightly anticorrelated with previous years
summer minimum ice area (R -0.21)
32What is the underlying trend?
Could take various fits to declining ice extent
September minimum ice extent - NSIDC
ice extent
33What is the underlying trend?
Could take various fits to declining ice extent
September minimum ice extent - NSIDC
ice extent
34What is the underlying trend?
Could take various fits to declining ice extent
September minimum ice extent - NSIDC
ice extent
But the trajectory of declining sea ice area is
inherently a nonlinear process that depends on
the survival rates
?
35What is the underlying trend?
Using trends in survival rates instead
survival rates from CICE hindcast
MY - ?
FY - ?
?
36What is the underlying trend?
Using trends in survival rates instead
survival rates from CICE hindcast
MY - ?
FY - ?
obs - NSIDC
CICE projection
ice extent
37IPCC AR4 Models 2007
Fig from RealClimate.org (2006)
38Summary
- Trend analysis
- sensitivity of September ice area to a changing
climate depends critically on the difference
between MY and FY ice survival rates - Perturbation analysis
- summer minimum area depends upon ice survival
rates in that year there is very little memory
of the previous years area - winter ice area does not significantly influence
summer area - About 20 of FY ice and 60 of MY ice survives
in CICE hindcast. - is this consistent with other models or
observations?
39Summary
- Trend analysis
- sensitivity of September ice area to a changing
climate depends critically on the difference
between MY and FY ice survival rates - Perturbation analysis
- summer minimum area depends upon ice survival
rates in that year there is very little memory
of the previous years area - winter ice area does not significantly influence
summer area - About 20 of FY ice and 60 of MY ice survives
in CICE hindcast. - is this consistent with other models or
observations?
Survival rate of MY ice in growth season
Survival rate of MY ice in melt season
40(No Transcript)
41Extra slides
42Loss of MY ice during growth season
43Latest CICE ModelV4 from LANL (developed by Hunke
and Lipscomb)
Tracer capability ice age, melt ponds, algae,
chemistry, etc Ice-thickness distribution Brine
pocket (mushy-layer) physics Elastic
viscous-plastic ice dynamics Incremental
remapping advection efficient for large numbers
of thickness categories and tracers, second-order
accurate in space, monotonicity preserving
State variables in each grid cell - 5 ice
thicknesses, 5 concentrations, 5 age classes, 5
snow depths, 25 temperatures, 25 salinities
(soon), and many other tracers No Atmosphere
(NCEP reanalysis), and Mixed Layer Ocean (OHF
from CCSM run, annually repeating)
441979-2006 March and September average ice extent
451979-2006 March and September average ice extent
obs
obs
106 km2
106 km2
R 0.60
R 0.83
cice
cice
106 km2
106 km2
46Predictability of 2008 summer minimum
2007 was a 3.5? year (assuming linear trend in
ice extent)
obs
47Predictability of 2008 summer minimum
2?S
2?S
95 confidence interval about linear trend
(assuming no memory of previous summers minimum
ice extent)
must have assumed incredibly low ice survival
rates!
obs
48Predictability of 2008 summer minimum
2?S
2?S
Need to look at trends and variability of FY and
MY ice survival rates from observations
49Larger trend in MY area than FY area
(f summer minimum FY ice area)
50Larger fractional change in MY area than FY area
(f summer minimum FY ice area)
(m summer minimum MY ice area)
51Larger fractional change in MY area than FY area
CICE hindcast
fractional FY area
fractional MY area
52Projections of future ice decline
projected ice area from fits to survival rates
from CICE hindcast CCSM coupled GCM SRES A1B
ensemble mean
year
53Stroeve et al (2007)
54Uncertainty in the decline of September ice area
55Uncertainty in the decline of September ice area
?
probability density
???? ?
56Analogy to uncertainty in climate sensitivity
Temperature
?
Roe (2007)
57Analogy to uncertainty in climate sensitivity
Temperature
Sea-ice area
?
?
Roe (2007)
58Uncertainty in the decline of September ice area
probability density
probability density
59September Ice Volume
Define a basin-average ice volume in terms of the
average thickness of ice within FY and MY
categories, and , respectively
Equilibrium Ice Volume
60Larger fractional change in volume than area
Define a basin-average ice volume in terms of the
average thickness of ice within FY and MY
categories, and , respectively
Equilibrium Ice Volume
Equilibrium Ice Thickness
61Larger fractional change in volume than area
Define a basin-average ice volume in terms of the
average thickness of ice within FY and MY
categories, and , respectively
Equilibrium Ice Volume
Equilibrium Ice Thickness
62Larger fractional change in volume than area
CICE hindcast
63Larger fractional change in volume than area
CICE hindcast
CCSM coupled GCM, SRES A1B ensemble mean
64Perturbation Analysis
65Perturbation Analysis
66Perturbation Analysis
67Perturbation Analysis
106 km2
106 km2
R -0.29
R 0.72
106 km2
106 km2
68Perturbation Analysis
106 km2
106 km2
R 0.96
R 0.50
69Perturbation Analysis
not significantly correlated winter ice area has
small influence on summer minimum FY ice area
106 km2
R -0.23
106 km2