Sea Ice, Climate Change and RemoteSea Ice, Climate Change and Remote
SensingSensing
Prof. David Barber
Canada Research Chair in Arctic System Science
Director, Centre for Earth Observation Science
University of Manitoba
Winnipeg, MB. Canada
www.umanitoba.ca/ceos
ESA Summer School, August, 2006
CEOSCEOS
Lecture outline
1) Arctic Climate Change and Remote Sensing
2) Thermodynamics, Geophysics and AOP/IOPs
3) Dielectrics, scattering and emission modeling
CEOSCEOS
Outline of this talk
• A look at thermodynamic processes
• Snow geophysics
• Sea ice geophysics
• A look at complexity
• Conclusions
CEOSCEOS
CEOSCEOS
CEOSCEOS
Complex Dielectric
!*=!"+j!#
Ice Type
Ice Thickness
Ice Salinity
Ice Temperature
Snow Depth
.
.
Multi-
frequency
& polarized
EM
Signatures
Forward Approach
Inverse Approach
Freeze onset date Radiative
transfer model
The electromagnetic properties of sea ice
IEEE TGARS, ONR ARI special issue. 36(5): 1750-1763
CEOSCEOS
Frequency/Polarization and
Sensor Geometry
!°total =
!°ss
+ "as
(#) *
!°sv
(#') + "s(#')
* !°
is + "
si (#") * !°
iv (#")
Snow Surface
Snow Volume
Ice Surface
Ice Volume
!h and L
Ri, Rw, $s, Wv, Ss, %s,
!h and L
Ri, Ra, Rb, $i, Is,
FrequencyPolarization
# of LooksIncidence Angle
CEOSCEOS
Thermodynamic Processes
CEOSCEOS
snow
Multilayer thermodynamic model
!
"scs
#Ts
#t=#
#zks
#Ts
#z+ I
o
$
% &
'
( )
!
"ici
#Ti
#t=#
#zki
#$i
#z+ I
o
%
& '
(
) *
!
"k( i / s)
#T( i / s)
#z+ (1" $
( i / s) )QSW +QLW " %& (i / s)TAI4"Qs "Ql " Io = L
(i / s)WAI
!
ks
"Ts
"zSI
= ki
"Ti
"zSI
!
"ki#Ti
#z IO
= Qw " LiWIO
ice
Snow/ice
Ice/ocean
Air/snow
Coupled column model
CEOSCEOSBarber et al. 1998
0-5-10
20
0
30
60
90
Depth
(cm
)
-5-10-15-20 -5-10-15 -5-10-15-20-20
Temperature (°C)
-15-5-10-15-20
Average T°
Diurnal !
Show
twave
Flu
x (
K ) 400
200
0
Snow
Sea Ice
5
Winter ablation 1 ablation 2 ablation 3 ablation 4 ablation 5
0-5-10 -15
Temperature is the control
Ice site overview, measurements
Tsnow Tice+water
Tair+Rh
U,V
u’,v’,w’,CO2’,H2O’SW PAR LW
SW LW
Q*
Tsurf
M1234
& T34
Ice sampling
T12
Spectral albedos
and
irradiance profiles
Temperature,
salinity, O-18,
for snow and ice
Nondestructive
transmittance
measurements
and irradiance
profiles.
CEOS
CEOSCEOS
Radiation, Heat and Mass Transfer Processes
of Snow over FYI
snow
ice
QH QELd LuKd Ku
K*o
K*B=Q*is
K*1x=Q*1x
Qso
Qs1x
QsB
Qio
Z=1x
Z=2x
Z=Nx
Z=0x
Z=B
QM
Q* Q*
Density (kg m-3)
Salinity (ppt)
Liquid (% by vol.)
Mass
CEOSCEOS
Snow
CEOSCEOS
0.0
6.0
9.0
12.0
15.0
18.0
Snow
Depth
(!s;
cm
)
10.0 20.0 30.0 40.0
Snow Grain Size (mm-2)
-8 -7 -6 -5 -4
Snow Temperature (Ts; °C)
0.00 0.02 0.04 0.06
Snow Brine Volume (Vb; %/100)
0.0 100 200 300 400 500
Snow Density ("s; Kg·m-3)
Grain Size3.0
Snow-Ice
Air-Snow
"s
#°
Vb
• Variables
• Processes
CEOSCEOS
0 1Scale (mm)
CEOSCEOS
•Snow is a complex crystalline material which forms
from the condensation and sublimation of water
vapour onto a nucleating material.
•Upon deposition the dendritic structures (small
angular crystal pieces which make up the snow flake)
break into fragments (a process known as saltation).
•This saltation process quickly increases the density of
the snow as it is blown across the Arctic sea ice.
•As the dendrites age a process called sintering occurs
(i.e., bonds forming at the points of adjacent
dendrites).
•This process results in an equilibrium density for
snow of about 375 kg·m-3 for snow on Arctic sea
ice.
Dendrites
CEOSCEOS
0 5Scale (mm)
CEOSCEOS
•Under temperature gradient metamorphism there is
a transfer of mass along the temperature gradient.
•This process is typified by the sublimation at the
warm end of the snow grain, transfer along the
vapour pressure gradient, and a corresponding
phase change from vapour back to a solid at the
colder end of the snow grain.
•This process results in a predominantly elongated
crystal structure with the long axis parallel to the
direction of the vapour gradient.
•The metamorphic state which results from this
process is often called kinetic growth snow grain.
Kinetic
structures
CEOSCEOS
0 10Scale (mm)
CEOSCEOS
•When water in liquid phase is low (or absent) equi-
temperature metamorphosis will create larger
grains at the expense of smaller grains due to the
vapour pressures associated with the snow grain
shapes.
•This is the principal process associated with early
spring grain growth or snow ripening.
•When water in liquid phase increases large grains
will combine into polycrystalline aggregates.
•When adjacent equitemperature grains aggregate
large single grain entities result .
•This usually coincides with draining within the
snow pack.
Aggregate
Structures
CEOSCEOS
Snow
Sea Ice
CEOSCEOS
Snow
Sea Ice
0.0
6.0
9.0
12.0
15.0
18.0
Snow
Depth
(!s;
cm
)10.0 20.0 30.0 40.0
Snow Grain Size (mm-2)
-8 -7 -6 -5 -4
Snow Temperature (Ts; °C)
0.00 0.02 0.04 0.06
Snow Brine Volume (Vb; %/100)
0.0 100 200 300 400 500
Snow Density ("s; Kg·m-3)
Grain Size3.0
Snow-Ice
Air-Snow
"s
#°
Vb
CEOSCEOS
Sea Ice
CEOSCEOS
Brine flux
New Sea Ice
CEOSCEOS
CEOSCEOS
CEOSCEOS
Snow is
saline
(Eicken’s chapter in Thomas and Dieckmann (ed) 2004)
First year sea ice microstructure
CEOSCEOS
Ice microstructure, light nilas, Cape Bathurst polynya
Station 124A on Oct 26, 2003
1.5cm
6cm
8cm
5mm thick section
through transmitted
light
1mm thin section
between polarizing
sheets
CEOSCEOS
CEOSCEOS
Crystal structure (FYI, Franklin Bay, May 9, 2004)
• Sea ice has irregularcrystal boundaries
• Growth parallel to (0001)plane is favored
$ Geometric selection tovertical c-axis orientationwith depth.
• Sizes increase with depth(related to growth rate)
• C-axis alignment withcurrents
• Close to bottom of thickFYI irregular c-axisorientations observed(no explanation)
1 cm
2.5 cm
4 cm
19 cm
14 cm
9 cm44 cm
39 cm
34 cm
29 cm 24 cm bottom
CEOSCEOS
Close-ups with microscope
viewed from above
8cm
6cm
Brine channel
Brine tube
Brine pocket Cellular substructure
platelet
Interconnected tubes
CEOSCEOS
1.5cm
4cm
124F
200E
{2.5cm
Very thin ice examples
Calm conditions
Agitated conditions
CEOSCEOS
Multiyear sea ice microstructure
.
10 t
o 3
0 c
m
Bubble
s ra
nge
from
0.1
to 1
.0 m
m
{
MeltP
onds
Hummocks
!' = 2.5-
3.2!" = 0.0-
0.1
}Winebrenner et al. 1989
CEOSCEOSBarber et al. 1998
0-5-10
20
0
30
60
90
Depth
(cm
)
-5-10-15-20 -5-10-15 -5-10-15-20-20
Temperature (°C)
-15-5-10-15-20
Average T°
Diurnal !
Show
twave
Flu
x (
K ) 400
200
0
Snow
Sea Ice
5
Winter ablation 1 ablation 2 ablation 3 ablation 4 ablation 5
0-5-10 -15
Temperature is the control
CEOSCEOS
.
MgCl2.12H2O
1000
500
100
50
10
5
1-10 -20 -30 -40 -50 -600
CaCO3.6H2O
Na2SO4.10H2O
MgCl2.8H2O
NaCl.2H2O
MgCl2.12H2O
CaCO3.6H2O
Na2SO4.10H2O
MgCl2.8H2O?
KCl
Na+
Cl-
SO4-
H2O
Mg+++Ca++K++rest
Cl-
H2O
TEMPERATURE (°C)
WE
IGH
T R
AT
IO (
g/k
g)
Ice
Bri
ne
NaCl.2H2O
KCl
Bri
ne
Ice
Sa
lts
Temp effect on the partial fractions of brine/ice and air
Phase diagram of sea ice showing the relationships between ice in solid phase, brine and
solid salts. After Assur, 1958.
CEOSCEOS
Some effects of temperature
Solid salts precipitate in brine
pockets (Light et al., 2002)
Brine tubes become
fragmented when
cooled (CASES, 10
Apr, 2004)
1m
m
Freezing Warming
Brine clusters merge
(CASES, May 16,
2004)
Gas bubbles
form within
tubes.
(CASES,
May 16)
CEOSCEOS
. .
Winter EarlyMelt
MeltOnset
AdvancedMelt
Dec April May/June June/July(Approximate time frames)
Brine Volume
Electrical Properties
Ice Strength PropertiesD
B
A Thermodynamics
C
Dep
th (
cm)
0
1.0
Str
eng
th (
MP
a)
12
34
1 Vertical Tensile2 Horizontal Tensile
4 Flexural
5 Compressive
3 Shear
5
2.0
70
0.1
0.4
Per
mit
tiv
ity
(
e ' )
Lo
ss (
e''
)
Ice Surface
Basal Snow Layer
Bulk Ice Volume
1
2
3
1.5
Permittivity
Loss12
0
-50
30
-10 -5 -2.5-7.5-20
Temperature oC
Amax
Bmin
Cmin
Cm
Cma x D
min
Snow
-100
Bm
Bmax
0
100
200
Bri
ne
Vo
lum
e (p
pt)
12
3
2.5
Dm
40
Basal Snow Layer
1
2
Ice
Am
Amin
Barber et al. 1996
The Rule of 5’s
CEOSCEOS
CEOSCEOS
Processes leading to melt pond
maturity over FYI. The melt
pond – albedo feedback is
initiated by an increase in the
atmospheric heat flux (L),
stimulating snow ablation and
the development of melt ponds.
During initial pond formation
the albedo of young ponds is
dictated by pond depth and the
scattering properties of the
frazil ice layer. Ablation of the
frazil layer is a function of the
wind induced mechanical
weathering, and solar
insolation (Wm-2). As the melt
ponds matured, the albedo is
dictated by pond depth and the
optical properties of the
columnar ice volume
CEOSCEOS
The Effect of Frazil Ice On Albedo
• The evolution of melt ponds through the melt season.Bubble densities quickly reduced after initial pondinghowever a ring of high bubble density is found alongpond fringes during times of advance.
CEOSCEOS
Pond Onset
CEOSCEOS
Pond Development
CEOSCEOS
Mature Ponds
CEOSCEOS
Pond Drainage
CEOSCEOS
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
165 168 171 174 177 180 183 186 189 192 195
Year Day
Sig
ma N
aught (d
B)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Alb
edo/P
ond F
raction
_º Albedo Pond Fraction 2 per. Mov. Avg. (_º)
a b c d e
Time Series !, "º and Pond Fraction
IOP and AOP of the seasonal ice cover
29.04.2005
Melting new snow
Early melt pond with 1cm of water
New snow collecting at rougher surface
Measurements - Spectral albedo site M3
CEOS
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
350 450 550 650 750 850 950 1050
m1_0407
m1_0411
m1_0412
m1_0415
m1_0418
m1_0422
m1_0424
m1_04290
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
350 450 550 650 750 850 950 1050
m2_0407
m2_0411
m2_0412
m2_0418
m2_0422
m2_0424
m2_0429
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
350 450 550 650 750 850 950 1050
m3_0411
m3_0412
m3_0415
m3_0418
m3_0422
m3_0424
m3_04290
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
350 450 550 650 750 850 950 1050
m4_0411
m4_0412
m4_0415
m4_0418
m4_0422
m4_0424
m4_0429
Spectral albedos from Button Bay, HB2005, April 7 to April 29, M-sites
M1 – small bare ice area M2 – hummock with ~15cm soft surface layer
M4 – hummock with ~5cm soft surface layerM3 – large bare ice area – early melt pond
1cm of water
on ice surface
resulted in lowest
observed albedo
Hummock (or white ice) sites
had clearly smaller variability
than bare ice (or blue ice).
CEOSCEOS
Spectral irradiance profiles and transmittance
0
20
40
60
80
100
120
140
160
180
200
220
240
260
0.1 1 10 100
650nm
600nm
550nm
500nm
450nm
400nm
CEOS
0
1
2
3
4
5
6
7
350 400 450 500 550 600 650 700 750
200
180
160
150
130
100
80
60
40
30
Diffuse attenuation
coefficient Kd(%) calculated
from irradiance profile
T1_0421 (hummock)
dep
th
water
ice
What about spatial variability?17 April 1998
25 June 1998
4 August 1998
What about AOP’s and IOP’s
at the bottom of the ice?
CEOS
Vertical profiles in sea ice for (a) temperature
(daily mean), (b) salinity and (c) density, with (d)
corresponding calculations of brine and air
inclusion volume fractions. Note the cut in scale
on the latter. The emphasis is on the bottom part
where high temperature and salinities resulted in
an off the scale increase in volume fractions.
Horizontal microstructure sections of a sea ice
sample taken on 9 May. The numbers on the
right-hand corner of each image indicate the
height above the ice-water interface from which
the section was extracted. Processing by:
1. Edge detect
2. Torn edges
CEOS
(a) The vertical downwelling irradiance profiles
integrated over PAR wavelengths and normalized
to bottom irradiance values. The corresponding
diffuse downwelling irradiance attenuation spectra
for the (b) interior ice and (c) bottom 10-cm
bottom layer, with comparisons to (d) particulate
absorption coefficient.
Average chlorophyll-a
concentrations measured on four
occasions using three different
methods to extract samples; ice
core drilled from surface (core),
4-cm thick ice puck taken by
diver from below (puck), and
bottommost algae layer sampled
by diver using syringe “slurp gun”
CEOSCEOS
A brief look at Complexity?
Snow
CEOSCEOS
Sea Ice Modeled Processes
snow
ice
atmosphere
ocean (mixed layer)
T = -1.8°C if ice exists
Kd Ld
Fa
Lu Ku Fs Fl
T = sfc temp
T = -1.8 C
Q*
Flux(1)
Flux(2)
Kd = downwelling SW flux
Ku = upwelling SW flux
Fa = absorbed SW flux
Ld = downwelling LW flux
Lu = upwelling LW flux
Fs = sensible heat flux
Fl = latent heat flux
Q* = net sfc flux
Flux(1)=snow-ice conductive
flux
Flux(2)=ice-ocean conductive
flux
Multiple layers (49)
Sfc nrg balance
Conductive fluxes
CEOSCEOS
80
75
70
65
605 55 105 155 205 255 305 355
Duration of OpenWater (Days)
Anomaly Position (Day of Year)
Role of snow in complexity
Effect of Moving a 5 day (20cm) Snowfall
Anomaly on Open Water Duration
(model run 1961-1990).
CEOSCEOS
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
107
110
113
116
119
122
125
128
131
134
137
140
143
146
149
152
155
158
161
164
167
170
173
176
Julian Day
Alb
ed
o
albedo(obs)
albedo(model)
Modeled vs. Observed Albedo
CEOSCEOS
Hourly vs. Daily Forcing
0
0.5
1
1.5
2
2.5
107
127
147
167
187
207
227
247
267
287
307
327
347 2
22
42
62 82
102
122
142
162
Julian Day
Th
ickn
ess (
m)
Ice (day)
Snow (day)
Ice (hr)
Snow (hr)
Ice
Snow
BU = 202
(July 21)BU = 189
(July 8)
FU = 289
(Oct 16)FU = 297
(Oct 24)
CIS Data: BU = July 9-16
FU = Oct. 15-22
Simulation: April 17, 1992 - June 19, 1993
Open water duration
difference = 21 days
CEOSCEOS
Bias of ‘land based’ versus ‘on ice’ forcing
0
0.1
0.2
0.3
0.4
0.5
0.6
107
111
115
119
123
127
131
135
139
143
147
151
155
159
163
167
171
175
179
Julian Day
Th
ickn
ess (
m)
Snow (day)
Snow (CONT)
Snow (SFS)
Snow (obs)
Q* Q* Tsfc Tsfc
CONT/obs SFS/obs CONT/obs SFS/obs
R-Square 0.44 0.56 0.91 0.98
Mean Error -2.5 -2 2.5 -0.08
St. Dev. 30 25 3.1 1.7
SFS more realistic snow & ice ablationHanesiak et al.
Conclusions
CEOSCEOS
Conclusions
1. Need to know geophysics and thermodynamics to determine
scattering and response to forcing
2. Dynamic vs Thermodynamic processes are NB
3. Many feedbacks exist and processes are not yet well understood
(and thus not modelled).
4. System is very sensitive to changes in snow thickness, distribution
and deposition (timing of sea ice formation is critical)
5. Assumptions of current processes applicable to the future