Southern Ocean Variability and Climate Change in HadCM3

1
Southern Ocean Variability and Climate Change in
HadCM3
Sheila Stark
2
Overview

• Introduction to HadCM3
• Southern Ocean Simulation water masses and
  ACC.
  
• SAMW / AAIW variability and comparison to
  observations.
  
• Future Scenarios.

3
HadCM3

• Coupled ocean atmosphere model without flux
  adjustment.
  
• Ocean component 1.25 x 1.25 with 20 vertical
  levels.
  
• Control integration (CTL) initialised with the
  ocean state of Levitus and Boyer (1994) and an
  1860 atmospheric composition (pCO2290ppm).
  
• Freshwater budget takes 370 years to come into
  balance.

4
Antarctic Circumpolar Current

• In common with many coupled climate models, the
  Antarctic Circumpolar Current (ACC) in HadCM3 is
  too strong
  
• Observations of the Drake Passage transport put
  it somewhere between 100 and 150 Sv
  
• HadCM3

5
Reasons for ACC drift

• Initial transport on large side (bathymetry
  important?)
  
• The ACC transport drifts upwards because the
  density structure changes over time
  
• This is mainly caused by a drift in salinity (on
  average too salty to the south and too fresh to
  the north)

Levitus
HadCM3
6
Water Masses SR3 Section
AAIW
LCDW
AABW
SAMW
7
AABW tracer results
Tracer added and continually replenished in the
surface layer for 5 years.
8
AABW tracer results 2
9
SAMW in the Indian Ocean

• 32S a well observed section occupied 3 times
  since 1962, and partially sampled as part of
  WOCE.
  

10
Forced Ensembles
Forced with historical records of volcanic
emissions and solar irradiance.
Forced with anthropogenic changes in
GHGs, ozone and sulphur.
Both natural and anthropogenic forcing applied.
11
Model Internal Variability

18 years
11 years
9 years
12
SAMW in HadGEM1
SAMW denser than in HadCM3 26.0 - 26.5kg.m-3.
Again the response is dominated by a freshening
trend, which is stronger than that seen in ANY of
the HadCM3 forced experiments.
HadCM3 trend driven by Freshwater Heat Wind
stress / Ekman
13
The 21st Century
-

core SAMW isopycnal
Long term freshening trend decadal fluctuations
14
What drives these changes?
Heat dominates the 21st century isopycnal
freshening
15
Caveats

• It is likely that HadCM3 overestimates the
  isopycnal freshening trend because
  
• Precipitation overestimated.
• Westerly winds too strong excessive Ekman
  contribution.
  
• Dominance of heat lends some credibility to the
  result, magnitude in doubt.

16
AAIW
Lighter (and shallower) than in observations.
Fresher and warmer than observed.
17
AAIW in the 21st Century
18
Model AAIW Variability
50 years
25 years
Longer timescales more dominant than in SAMW
19
Summary 1

• Models useful as the Southern Ocean is
  relatively data sparse.
  
• Model simulation reasonable
• ACC too strong.
• Deep water masses not simulated well.
• Thermocline water masses too light and fresh.
• How important is the simulation for climate
  change response / heat uptake?
  
• Observed changes in SAMW are not detectable
  ?internal variability.

20
Summary 2

• Have to assume that the model internal
  variability is realistic more observations are
  needed
  
• Model predicts a long term freshening of both
  SAMW and AAIW reflecting heat uptake and
  increasing high latitude precipitation.