What can inadequate observations

1
What can inadequate observations tell us about
incomplete models?
Steve Rintoul Antarctic Climate and Ecosystems
Cooperative Research Centre CSIRO Marine and
Atmospheric Research CSIRO Wealth from Oceans
National Research Flagship
2
Outline

• Southern Ocean role in climate
• What do we know from observations?
• Antarctic Circumpolar Current (ACC)
• Overturning circulation
• Is the Southern Ocean changing?
• Open questions

3
Southern Ocean role in climate
4
Sabine et al., 2004
5
More than 90 of global warming since 1955 is
found in the ocean.
Levitus et al., 2004
6
Most of the change in ocean heat content since
1955 is in the southern oceans.
Levitus et al., 2004
7
What we would like to know

• Absolute velocity/property transports
• Mixing
• Eddy fluxes
• Water mass formation rates
• Sea ice (including volume)
• Forcing
• Mean and variability of above

8
What we cant observe

• Absolute velocity/property transports
• Mixing
• Eddy fluxes
• Water mass formation rates
• Sea ice (including volume)
• Forcing
• Mean and variability of above

9
Antarctic Circumpolar Current
10
WOCE SO Atlas, Orsi and Whitworth
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Rintoul and Sokolov, 2001 Cunningham et al.,
2003
15
Baroclinic transport of the ACC
6 repeats of WOCE SR3 Mean 147?10 Sv Heat
transport varies by 0.6 x 1015 W (relative
to 0?C). ACC ? interbasin exchange
Rintoul and Sokolov, JGR, 2001
16
ACC transport in neutral density layers Australia
(SR3) color Drake Passage (SR1) black
Rintoul and Sokolov, 2001 Cunningham et al.,
JGR, 2002
17
What sets the transport of the ACC?
18
warm
cold
Low-latitude closure matters too (diffusion?
outcrop in NH?), and properties of inflow.
Flow topography interactions alter structure of
mean flow and eddy fluxes. Transport becomes a
complicated function of forcing, eddies,
topography (and hence model details).
19
ACC transport in high res. models
Grezio et al., 2005
20
Zonally-integrated momentum balance
-fV1 - ?'1p'1x ?o - R1
?1
Surface (includes Ekman)
-fV2 ?'1p'1x - ?'2p'2x - R2
?2
unblocked by topography
?3
-fV3 ?'2p'2x - hpbx - R3
blocked layer
V net meridional volume flux ?o wind stress ?
layer thickness p pressure R Reynolds
stress divergence pb bottom pressure
Olbers
21
Is the ACC in Sverdrup balance?
ß?x ?pb??H ??? ??F
Bottom pressure torque (color) barotropic
streamfn (black)
Rintoul, Hughes and Olbers 2001
22
How variable is ACC transport?

• Barotropic ? ?8 Sv
• Meredith et al., Drake bottom pressure gauges
• Baroclinic ? ?10 Sv
• 8.5 Sv, Cunningham et al., 2003 Drake Passage
• 9.5 Sv, RintoulSokolov, 2001 note this is
  interbasin exchange, includes westward flow south
  of Tasmania
• 6.1 Sv, Rintoul et al., 2002 altimeter data,
  south of Tasmania
• BT dominates on short time-scales (lt1 yr?), BC on
  long time-scales
• Free BT mode in south
• eg Hughes et al., 1999, 2003 Vivier et al., 2005

23
Southern Annular Mode/Antarctic Oscillation
Thompson and Solomon, Science, 2002
24
ACC response to the SAM

• Positive SAM causes an increase in ACC
  transport. True or false?
• Hall and Visbeck ?SAM of 1.5 ? 1 Sv
• Meredith et al. 2 ? 2 Sv
• Oke and England 5 degree poleward shift of
  winds results in 5.7 Sv increase in ACC

25
ACC response to climate change

• Greenhouse warming causes an increase in ACC
  transport. True or false?
• Bi et al., 2002 takes 150 years to start
  increasing. Increase of 16 Sv after 300 years.
• Saenko et al., 2005 increase of 27 Sv after
  equilibrium with 4xCO2 forcing
• Fyfe and Saenko, 2005 increase of 16 Sv in 100
  years.

26
Frontal structure of the ACC
Sokolov and Rintoul, in prep.
27
Values of absolute sea-surface height which
coincide with the largest number of points with
sea surface height gradients above a threshold
value. We find the distribution of high
gradients in SSH is strongly peaked at particular
values of SSH. ? We can track fronts using
contours of absolute SSH.
28
(No Transcript)
29
Belkin and Gordon Orsi et al. Moore and Abbott
Sokolov and Rintoul, in prep.
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
ACC Summary

• ACC consists of multiple, robust filaments.
• Eddies carry momentum downward and both heat and
  mass poleward.
• ACC and overturning circulation are intimately
  linked. Wind and buoyancy forcing, eddy fluxes,
  mixing, and topographic interactions all
  contribute to the dynamical balance.
• ACC transport is insensitive to changes in the
  wind?
• Divergence of ACC property transport (eg heat
  flux) is the important quantity for climate.
• Modeled ACC transport is sensitive to model
  details - not a very good metric for ocean
  climate models?

34
(No Transcript)
35
(No Transcript)
36
The ocean is cold.
http//www.ewoce.org
37
Diapycnal spread of a tracer at 300m Kv0.17 x
10-4 m2s-1 Ledwell et al. 1998
Mixing is too weak to support enough upwelling to
balance sources of dense water.
38
Spatial distribution of Kr in the SO
log10 (Kr m2 s-1) along ALBATROSS cruise track
(Naveira et al., 2003)
39
Air-sea buoyancy fluxes drive water mass
transformations in the Southern Ocean which
close the loop of the global overturning
circulation cells.

Speer, Rintoul and Sloyan, JPO, 2000
40
Impact of high-southern-latitude heat gain
Air-sea transformation (Sv)
Consuming deep water
Forming mode int. water
Speer et al., 2000, Sloyan and Rintoul, 2001
(UWM/COADS) Karstensen and Quadfasel, 2002
(UWM/COADS, NCEP/NCAR)
41
Lumpkin and Speer, 2005
42
Sarmiento et al., 2004
43
(No Transcript)
44
Ekman eddies (residual mean)
Ekman pumping
Southern Ocean buoyancy flux determines upwelling
pattern, hence nutrient supply and community
structure.
Ito et al., 2005
45
SO Overturning Summary

• Water mass transformations in the Southern Ocean
  connect the lower and upper limbs of the global
  overturning circulation.
• Southern Ocean therefore plays a primary role in
  global heat, freshwater, carbon, and nutrient
  budgets.
• Estimates of the overturning strength from
  observations continue to span a wide range.

46
Is the Southern Ocean changing?
47
Intermediate depth waters in both hemispheres
have become fresher in recent decades.
Wong et al., 1999
48
Curry et al., Nature, 2003
49
Observations south of Australia show large
variability in mode water properties from
year-to-year, driven by changes in cross-frontal
Ekman transport (not air-sea fluxes).
Circles show T-S properties of SAMW south of
Tasmania size of dot is proportional to strength
of mode. Triangles and squares are data from 1968
and 1978.
Rintoul and England, JPO, 2002
50
Floats reveal the Southern Ocean is warming
faster than global average rate.
Gille, 2002
51
Southern Annular Mode/Antarctic Oscillation
Thompson and Solomon, Science, 2002
52
Summary of ocean response to SAM

• Southward shift and strengthening of winds drives
  stronger northward Ekman transport and upwelling
  of deep water.
• ACC shifts south and transport increases
  slightly.
• Likely impact on sea ice (but what is it?)
• More upwelling of relatively warm water may
  contribute to increase in glacial ice melt and BW
  freshening (speculation).

Hall and Visbeck Meredith et al Oke and
England Sen Gupta and England
53
(No Transcript)
54
Freshening of Ross Sea shelf waters
Jacobs et al., Science, 2002
55
(No Transcript)
56
0.017 psu
57
.015
.008
.009
58
depth
depth
?
?
Distance above bottom (0-400 m)
salinity
salinity
?n
?n
2005
1995
59
(No Transcript)
60
(No Transcript)
61
140E (SR3) Adelie Land BW
1968-1971
2003
2002
Aoki et al., submitted
62
1968-1971 summer
63
WOCE, summer
64
Dickson et al., 2002
65
0.012 psu/decade
Dickson et al., 2002
66
Causes of fresher shelf water

• Increased glacial ice melt?
• More precipitation?
• Less sea ice formation?
• Change in winds and ocean circulation?

Davis et al., Vaughan Science, 2005
67
Summary of AABW changes

• Bottom water throughout the Australian-Antarctic
  Basin is fresher, lighter and higher in oxygen in
  2005 than in 1995.
• The recent changes continue a trend extending
  back to 1970, resulting in a dramatic basin-wide
  shift in the T-S curve.
• Rate of change is comparable to North Atlantic
  and may be increasing. Aliasing?
• Both Adelie Land and Ross Sea bottom water
  sources have become fresher.

68
Where models can help

• Interpolate between sparse observations to
  characterise variability
• Extrapolate (predict future)
• Diagnose dynamics and climate mechanisms
• Experiments (what if ?)
• Role of eddies
• Barotropic flow

69
Questions

• (Why) is the ACC transport relatively insensitive
  to changes in wind?
• Topography?
• Eddies?
• Can coarse resolution models capture correct
  response of ACC to change in forcing (when both
  flow and topography are smoothed)?

70
Questions

• What is the magnitude of the residual circulation
  today and how will it change?
• To what extent do eddies compensate Ekman
  transport?
• What is the vertical and horizontal distribution
  of eddy mass transport?
• Do present eddy parameterisations (GM) do the
  right thing?
• Whats ?(x,z)?

71
Questions

• What has caused observed changes in the Southern
  Ocean?
• SAM?
• Climate change?
• Natural variability?

72
Questions

• Is diapycnal mixing elevated along the path of
  the ACC? Is it important?
• Is there any evidence of feedback from Southern
  Ocean SST to the atmosphere and regional climate
  in the SH? From sea ice anomalies?

73
Recommendations

• Water mass distributions and baroclinic flow are
  about the only thing we measure well.
• Luckily, they provide a sensitive and relevant
  test for a climate model to simulate climate
  change and variability, the model must capture
  the formation and subduction of water masses
  correctly and get mean state right.
• Unluckily, getting this right in climate models
  is not so easy.

74
Recommendations

• Observational evidence of change is still
  fragmentary, but growing. This evidence is
  another powerful test for models, a test they may
  be beginning to pass.

75
Recommendations

• Declare a moratorium on observations in the North
  Atlantic for a decade and divert the resources to
  the Southern Ocean.

76
(No Transcript)
77
(No Transcript)
78
(No Transcript)