Understanding Paleoclimates Modelling the GlacialInterglacial Climate with Coupled GCM Ayako AbeOuch

1
Understanding Paleoclimates Modelling the
Glacial/Interglacial Climate with Coupled GCM
Ayako Abe-Ouchi, CCSR, University of Tokyo /
FRSGC

• 1 Introduction
• 2 Model description
• 3 Greening Sahara (mid Holocene)
• 4 Glacial Ocean
• 5 Ice sheet evolution
• 6 Summary

2
Introduction

• Earth System Modelling for Climate in the Past
  (Paleoclimate)
• --- To Understand the climate behaviour
• --- To validate the GCM that we use for Future
  prediction
• Model for Paleoclimate
• ---gt 1 needs long integration
• ---gt 2 needs feedback loops among several
  subsystems.
• ---gt 3 needs high resolution if the phenomenon is
  regional.
• Preliminary results are presented

3
Model description

• Atmosphere CCSR/NIES/FRSGC AGCM T42L20 (?simple
  EMBM)
• Ocean CCSR COCO 10.5 lat x 1.4lon, L43
• Sea ice Elastic Viscous Plastic model with 0
  layer thermodynamics.
• Ice Sheet Three dimensional thermo-mechanical
  coupled model
• Dynamical Vegetation to be coupled (LPJ and
  Kissme)
• Carbon Cycle to be coupled

4
16
Surface Temperature in Coupled GCM
SST Model -Obs.
18
Air Temperature (2m)
CO2 1/yr
16
?
Model After 70 years
SST Model -Obs.
CT02605
14
Observation
Low-Mid latidude drift occurs in the first few
years (fast initial response) High latitude
drift appears after the initial drift.
CT02603
Model After 70 years
year
5
Precipitation
Obs.(CMAP) AGCM CGCM
DJF
DJF
JJA
6
Greening Sahara
Polen(Hoelzmann et al., 1998)
Savanah
Steppe

• Not enough sensitivity of model climate of AGCM
  only.
• ---gt PMIP2

7
Sea Surface Temp. Change and Monsoon at 6000 yr BP
8
Precipitation Change in Coupled GCM (zonal mean
20W-30E)
Coupled GCM Climatology affects The response of
the Rain belt.
9
Modification of Physical Processes in AGCM
Modification in AGCM physics moistenning the
troposphere improved the response
Chikira 2003
Trop. Rain Forest
Steppe
Biome (Prentice et al., 1992)
Savanna
Dessert
10
4. Glacial Ocean and climate

• Response of surface ocean and thermohaline
  circulation to external condition is of interest.
  Ice age climate can be checked by rich data.
• Carbon cycle which involves the surface and deep
  sea could be related to the low CO2.
• Ice Age data show a large climate variabilty.
• (Modification of ENSO be discussed.)
• Without flux adjustment and some spin-up
  technique
• Control Experiment
• vs.
• Low CO2 Experiment
• Is conducted.

11
Time series of Thermohaline circulation
Control CO2x2 Glacial CO2
12
Overturning (THC) in the North Atlantic
9
11
14
10
Glacial
Control
8
7
Antarctic Bottom Water dominating more in the
Glacial than the Control.
13
Ocean Heat Transport
Present (Red line) vs. Glacial World (Blue dashed
line)
More heat to the south and less heat to the North
at the Glacial.
14
Formation of NADW and Sea Ice in North Atlantic
Modern
Glacial
Winter (Feb.) Convection in the north of
Iceland disappears in Glacial Ocean
15
Formation of NADW and Sea Ice in North Atlantic
(2)
Modern
Glacial
Future Warming
Winter (Feb.) Convection in the north of
Iceland disappears in Glacial Ocean
16
5. Ice Sheet Evolution

• Why did glacial/interglacial cycle of 100 ka
  cycle occur?
• ---gt Oscillator of 100 ka? (CO2, eccentricity.)
• ---gt cc. 20 ka, 40 ka oscillator - resonance or
  nonlinearity of the system?
• ---gt 100ka forcing phase locked some oscillator

Wavelet Analysis (Hargreaves and Abe-Ouchi 2003)
17
Phase diagram of a simple model response to
periodic forcing of 20ka(Abe-Ouchi, 1995)
18
?????????????
Ice Sheet Model in ESM
CCSR/NIES/ AGCM
monthly mean Temperature and Precipitation
Ice thickness, Bedrock sinking Ice temp. and flow
3D thermo-mechanical ice sheet model (Saito
and Abe-Ouchi, 2002) Shallow ice
approximation Thermodynamics-dynamics
coupling Simple sliding applied Bedrock isostacy
included Horizonal resolution 1 deg
lon./lat. Vertical 20 layers Degree Day mass
balance model
19
Temp.(JJA) and Net Mass Balance of LGM
Snowfall/ Sublim ./ Melt / Net (mm/year)
Temp.( LGM-Present) Annual mean -17.8K JJA
-21.1K Net Mass Balance over Laurentide Ice
sheet
20
Ice sheet - atmosphere feedback

• Ice albedo feedback
• Elevation - mass balance feedback
• Stationary wave feedback (through temperature)
  (Cook and Held, 1988)
• Transient eddy feedback (through precipitation)
  (Hall et al, 1990, Kageyama and Valdes, 1997)

21
Cooling due to Ice Sheet Existence Temperature
drop (K) Control minus LGM
Total cooling
Albedo Effect
Lapse rate effect
Residual
22
Stationary eddy- Temperature feedback (Full -
Flat ice LGM assuming lapse rate of 5 K/ km)
500hPa Z(m) (JJA)
Temperature change (K) At 850hPa (JJA)
23
Ice Sheet Modelling using the AGCM results at LGM

• No Ice LGM
• without ice sheet feedbacks

(b) Albedo effect But no lapse rate effect
(c) Only lapse rate effect
(d) Albedo effect lapse rate effect
(e) Full LGM
24
Dependence of net ice mass balance on ice sheet
size
LGM 21 ka
Ice Cap
Ice Sheet Size
18 ka
15 ka
No Ice
12 ka
Forcing
LGM
25
Dependence of ice sheet budget on forcing CO2
vs orbital
Grill the
Grill the LGM ice sheet by different
forcing Experiments with Cold vs Hot orbit
e0.05 Cool vs Warm orbit e 0.015 CO2 low 200
ppm Pre-Ind. 280 ppm high 345ppm
26
Dependence of net ice mass balance on CO2 and
orbital forcing
Dependence of net ice mass balance on CO2 and
orbital forcing
Cold orbit High CO2 -84.8
Cold orbit Low CO2 0
Hot orbit Low CO2 -259.0
Hot orbit High CO2 -436.0
27
Ice Sheet- Atm-Ocean coupling
Response to Orbital parameters (warm-cool orbit)
With Ice Sheet
Without Ice Sheet
Ice topo.
Air Temp. JJA
Ocean.
28
Summary

• Climate Change could be affected by the model
  control climate. Careful consideration of
  moisture process affects the whole paleoclimate
  discussion.
• ES enables the long term integration and a lot of
  experiments for the past climate.
• Different Hierarchy of models should be used. GCM
  could help the simpler model to identify the
  processes that should be included with higher
  priority.

29
Conclusion (2)

• Laurentide do help the Fennoscandian ice sheet to
  grow in the western part through the transient
  eddy feedback.
• Growth of Fennoscandinan ice sheet to the south
  in the western part is prevented by the
  stationary wave feedback of Laurentide ice sheet
  and the presence of itself.

30
Summary

• 1. Phase diagram of ice sheet response to
  periodic forcings of 20ka show that the100
  ka-like response occurs in a certain range of
  phase space of forcing.
• 2. Especially the summer maxima of this mode
  locates in a limited range, which corresponds to
  the area of multiple equilibria.
• 3. In case of Laurentide ice sheet, multiple
  equilibria seems to exist even under the LGM
  forcing. Threshold of ice sheet size/shape is
  between 15 and 18 ka ice size .
• 4. It is likely that the response time in this
  area of multiple equilibria can become very long
  under certain environmental condition, such as
  the climatic forcing and bedrock response.
• 5. The speed of growth and retreat of ice sheet
  could be highly dependent on the strength of
  feedbacks.
• 6. Orbital forcing may have a larger impact on
  ice sheet than CO2 even for 100ka cycle change.

31
Conceptual threshold model for the
glacial-interglacial cycles.

• The termination always following the smallest
  maxima in
• summer insolation but always follow the smallest
  maxima in
• A summer insolation

• A model able to switch abruptly
• between different climatic modes,
• in relation to both astronomical
• forcing and ice sheet evolution.
• Thresholds (for both insolation and
• Ice volume) and time constants are
• important.
• For each mode, the ice volume
• equation is linear,

(Paillard, 1998)
32
Simulation of NH ice volume under both the
insolation and CO2 change

• Successful simulation of ice volume by an EMIC.
  (2D- lat.and vertical)
• Sensitivity of NH ice to CO2 is not constant.
• Relative importance of CO2 vs. Orbit depends on
  model. (cf, Tarasov and Peltier, 1997)

Berger et al (1998), Li et al (1998)
33
This talk

• Here we focus first on a single oscillator as an
  example and show the possibility of producing
  100ka -like oscillation (longer than the one of
  forcing) by a realistic ice sheet model. Several
  sensitivity studies are performed also by GCM.
• Response of ice sheet to periodic forcing by a
  2-dim ice sheet model .
• The thresholds and response time in GCM for
  Laurentide ice sheet to understand the
  termination mechanism.

34
Temperature change (K)over Laurentide ice sheet
35
Topography Effect upon Cooling
From the exps. of Full LGM - Flat ice LGM run,
Lapse rate of 5 K/ km is estimated. Residual is
the component that Cannot be explained by the
change assuming the lapse rate.
Lapse rate effect
Residual
(K)
36
Precipitation change rate from Full, flat, no
ice LGM runs
(a) Full LGM
(b) Albedo Effect
(c) Topography Effect
37
Response of ice sheet to periodic forcing of 20
ka in a 2-D ice sheet model
38
Equilibria of ice sheet and the phase diagram
Range of summer maxima for chaotic response
39
Inception and Ice sheet growth
Ice sheet can initiate with the help of small
scale topography ( 50km size).
Abe-Ouchi and Blatter (1993)
40
Uniform 6K cooling With altitude albedo
feedback.
41
Effect of Strength of Feedbacks on ice sheet
evolution and Equilibria
Uniform 6K cooling With only altitude feedback.
Uniform 6K cooling With altitude albedo
feedback.
LGM forcing
Strength of the Albedo feedback controls both the
final equilibrium state and the speed reaching a
certain size (time constant).
42
Equilibria of ice sheet and the phase diagram
Range of summer maxima for chaotic response
43
Retreat speed
Retreat speed (Time constant?)
Retreat speed of the ice sheet is highly
dependent on the forcing change (a, b and c) and
the delay of bedrock response.
44
Threshold/ Critical Ice sheet size and shape
Around the threshold, the relative relation
between the current ice sheet size and the ice
sheet size at the threshold becomes critical. The
response time of ice sheet could be very large or
small.
Ice Cap
Ice Sheet Size
No Ice
Forcing
LGM
45
Impact of ice sheet size upon climate
Difference in summer air temperature
12ICE-LGM
15ICE-LGM
46
???????????????? ??????????????
47
Threshold/ Critical Forcing
Around the threshold, any small forcing can push
the ice sheet into a new mode.
Ice Sheet Size
Ice Cap
current
No Ice
Forcing
48
Dependence of net ice mass balance on CO2 and
orbital forcing
Dependence of net ice mass balance on CO2 and
orbital forcing(2)
Impact of Orbit parameters are
49
Summary

• Green Sahara
• Glacial Ocean
• Ice

50
(No Transcript)
51
Around the threshold the response time of
ice Sheet is very large.
52
Response of climate to orbital parameters
Suarez and Held (1978)
53
(No Transcript)
54
????????
55
Conceptual threshold model for the
glacial-interglacial cycles(Palliard, 1998)
Conceptual threshold model for the
glacial-interglacial cycles.(Palliard, 1998)

• The termination always follow the smallest maxima
  in summer insolation.
• Importance of thresholds and time constants for
  each mode.

56
(No Transcript)
57
(No Transcript)
58
Sahara??
59
Conclusion (2)

• Laurentide do help the Fennoscandian ice sheet to
  grow in the western part through the transient
  eddy feedback.
• Growth of Fennoscandinan ice sheet to the south
  in the western part is prevented by the
  stationary wave feedback of Laurentide ice sheet
  and the presence of itself.

60
????????
??????????
(?)??65?????? (?)?????

• ????????????????????????????????????????????????
  ?????????(Hays et al (1976))
• ????????????????????????????????????????(Imbrie
  et al(1993))

?????????????????
61
Mode of Glacial/Interglacial cycles and the role
of ice sheet

• Why did glacial/interglacial cycle of 100 ka
  cycle occur?
• ---gt Oscillator of 100 ka? (CO2, eccentricity.)
• ---gt cc. 20 ka, 40 ka oscillator produces nearly
  100ka cycle through some mechanism of resonance
  or nonlinearity of the system.
• ---gt 100ka forcing phase locked some oscillator