The 8.2Kyr event Julia Tindall Freshwater hosing experiments Ron Kahana

1
The 8.2Kyr eventJulia TindallFreshwater hosing
experiments Ron Kahana
2
The 8.2Kyr event

• Introduction and motivation
• The 8.2Kyr event in data
• Cause of the event
• Modelling the 8.2Ka event using other models
• Modelling the 8.2Ka event using HadCM3L

3
The 8.2Ka event in Greenland ice cores

• Largest rapid climate change
• event of the Holocene
• (cooling of 3oC-6oC)
• Useful for understanding
• the sensitivity of the climate
• and the likelihood of a similar
• future event
• The ideal test for climate
• models

4
Data from Greenland ice core Figure from Alley
and Ágústsdóttir 2005
5
The 8.2Kyr event globally?
6
Timing and structure in Greenland(d18O from
GRIP(red) and GISP2(black))
From Thomas et al 2006
7
Summary of evidence for 8.2Ka event globally

• Recent review (Morrill et al. 2005) found a
  statistically significant
• signal at 8.2Ka in 40 of records considered
  in both the Northern
• Hemisphere and the tropics
• Important to separate a clear 150yr 8.2Ka
  signal from
• millennial scale variability in the Holocene
• Was sharp 150year event superimposed on a
  longer (millennial scale)
• weaker event
• No evidence for event over Southern Hemisphere,
  or southward shift
• of ITCZ
• Some evidence of a slowdown in NADW formation
  at 8.2Ka,
• although this evidence is weak as many proxy
  records
• contain no signal

8
Cause of the 8.2Ka event
9
Details of outflow from Glacial Lake Agassiz

• 151,000km3 of freshwater
• 5.2Sv over 6months/1year
• Reasonably well dated and occurred at 8.45Ka

10

• Legrande et al 2006
• GISS (model E)
• Ensemble of experiments with 2.5Sv 5.0Sv added
  over 6 months to 1 year
• Large differences between ensemble members
• All ensemble members, had a full recovery of the
  THC within 30 years although sometimes there were
  secondary shutdowns.

Temperature
precip
d18O in precipitation
d18O in seawater
11
Modelling the 8.2Ka event using other models

• Wiersma et al 2006
• ECBilt-Clio model (intermediate complexity)
• Flood equivalent to 5.2Sv

With baseline flux of 0.172Sv
Without baseline flux
12
Other of previous modelling results

• NCAR model has full recovery in 10 years
  (Carrie Morill 8.2Ka workshop)
• Vellinga and Wood 2001 HadCM3 forced with 16Sv
  years recovery in 120years
• Bauer et al 2004 CLIMBER-2, multi-century weak
  freshwater pulse (0.04Sv) required (associated
  with melting of LIS)

13
Experiments with HadCM3
Expt 1 Expt2
Timing Instantaneously Over one year
Depth Top 800m Top 10m
Area Near Labrador Sea (84 gridboxes) N. Atlantic (50oN-70oN) (103 gridboxes)
d18O 0 -30
Other issues Spin up not completed
14
HadCM3 5Sv added over North Atlantic for 1
year First 10 years of model run show cooling
over much of the Northern hemisphere however
d18O signal is more noisy.
15
Atlantic MOC
16
(No Transcript)
17
First 10 years
Next 10 years
Last 20 years (yr 57-yr 77)
Temperature Changes
d18O Changes
18
What could improve model results?

• 8.2Ka boundary conditions
• Extra freshwater forcing
• (e.g. preflood0.055Sv, flood2.5Sv,
  routing0.172Sv
• rerouting0.104Sv ????)
• Other initial conditions

19
Summary

• 8.2Ka event is the largest rapid climate change
  event to have occurred in the Holocene.
• It is a good test for climate models and could
  provide information about the sensitivity of the
  climate and hence the likelihood of future rapid
  climate change
• Attributed to the final drainage of Lake Agassiz
  which released 5.2Sv of freshwater into the
  North Atlantic for 6months-1year at 8.45Ka
• Focus of a number of modelling studies, but often
  models need more than the suggested amount of
  water to produce a realistic response.
• Initial experiments with HadCM3 suggest that the
  model will be able to reproduce a realistic
  representation of the 8.2Ka event, although
  difficulties may occur with accurately
  representing the duration and the 250year lag.