North Pacific Decadal Variability: Role of Ocean-Atmosphere Coupling

1
North Pacific Decadal VariabilityRole of
Ocean-Atmosphere Coupling

• Lixin Wu and Zhengyu Liu
• Center For Climatic Research
• University of Wisconsin-Madison

Collaborators Dr. R. Jacob, Dr. R. Gallimore,
Ms. D. Lee and Ms. Y. Zhong
2
Paleoclimate Observation
Tree Ring Reconstruction

Southwest Drought
Northern Plains Drought
D. Stahle and E. Cook
3
Modes of Pacific Decadal Variability
Tropical Pacific Mode
North Pacific Mode
Eastern North Pacific Mode
4
Fundamentals of Pacific Decadal Variability

• Two Key issues
• Where does the memory reside?
• (2) What is role of the O/A feedback?

5
Potential Mechanisms

• Variability Process Mechanism

Teleconnection ??
6
Coupled Modeling Surgery

• Grand Geophysical Laboratory
• Coupled Climate model
• FOAM_1.5 (Fast Ocean Atmosphere Model,
  Jacob,1997)
• AGCM CCM2 Dynamics/CCM3 Physics (R15)
• OGCM POP-like (2.8o1.4o24-level)
• No flux adjustment
• Modeling Surgery
• Partial-Coupling constrain A/O coupling in
  selected regions by selected variables
• Partial-Blocking Constrain regional oceanic
  teleconnection

7
Model NPM and ENPM
Z500
SST
ENPM
NPM
Wu and Liu, 2003
8
North Pacific Decadal Variability
No ENSO
No coupling
9
NO ENSO
NO A/O Coupling
ENPM
ENPM
NPM
NPM
10
Modeling Evidence

• Two Distinctive Decadal Modes in the North
  Pacific
• NPM Coupled ocean-atmosphere mode
• ENPM Stochastic Mode under the influence of
  tropics in conjunction with oceanic Rossby wave
  propagation ( Spatial resonance)

11
Midlatitute SST-Atmosphere Feedback

• Observational Evidence?
• Target Unclear
• Modeling simulation?
• Diverse and confusing

12
Previous Model Simulations
Reference Exp. Resoln. Response
Palmer and Sun (1985) AMIP 350km15 EqBt high
Pitcher et al. (1988) AMIP R159 EqBt low (W and C)
Kushnir and Lau (1992) AMIP R159 EqBt low (W and C)
Ferranti et al. (1994) AMIP R6315 EqBt high
Peng et al. (1995) AMIP T4221 EqBt high (Nov.) Baroclinic low (Jan.)
Kushnir and Held (1996) AMIP R159 Barclinic low
Latif and Barnett (1995,96) AMIP T4219 EqBt high
Peng et al. (1997) AMIP T4219 EqBt high (Feb.) Baroclinic low (Jan.)
Yulaeva et al. (2001) FHXE T4219slab Baroclinic low
Sutton and Mathieu (2002) FHXE T4219slab Baroclinic low
13
Observational evidence inferred from
lagged correlation of SLP, surface wind and
Kuroshio Extension SST
Model
Model Ctrl
NCEP
Atm. lead
Ocn. lead
Ocn. lead
Atm. Lead

• Atmospheric forcing of ocean, dominant
• Atmospheric response to ocean, with season

Liu and Wu, 2004
14
Statistic EstimationFrankignoul et al., 1998
A(t) ? T(t) N(t) A atmosphere, T SST,
N atmospheric noise Since ltT(t-?), N(t)gt 0 lt
T(t-?), A(t)gt ? lt T(t-?), T(t)gt lt T(t-?),
N(t)gt ? lt A(t), T(t-?)gt/ltT(t),T(t-?)gt
15
Statistic Estimation of Atmospheric Response
NCEP
Surface wind
850 hPa ?
H
H
250 hPa ?
H
H
16
AMIP (Fixed SST forcing)
SST anomaly
17
Wet
Dry
AMIP Response
Z850
Z250
H
H
18
Coupled Ensemble Experiment
Initial value approach
Initial mixed Tgt0 (Nov.1st)
Atmosphere
Ocean
MLTA
Ensemble simulations (Dec. mean)
19
Wet
Dry
Fully Coupled Response
Z850
Z250
H
H
20
Thermally Coupled Ensemble Response
Dry
Wet
Z850
Z250
H
H
21
Geopotential Height
Temperature
Diabatic Heating
AMIP
?
FCE
TCE
22
Why weak heat flux but stronger warm-ridge
response in Fully Coupled Ensemble ?
23
Wet
Fixed Heat Flux Response
Z850
Z250
L
L
24
(No Transcript)
25
Modeling Evidence

• The atmospheric response depends not only on SST
  but also on flux boundary condition. The correct
  response can be attained only in fully coupled
  O/A system with a natural combination of SST and
  heat flux.

26
Why weak heat flux in Fully Coupled Ensemble ?
vTy
SST Heat Budget
Heat flux
Fully Coupled
Conv
Thermally Coupled
27
SST Propagation in FCE
Slow oceanic process
Fast Atmospheric Bridge
28
ATMOSPHERE
Model Climatology Prescribed
wind stress Anomaly
Heat and Moisture Flux
Wind stress
OCEAN
29
Coupled ocean-atmosphere response
WARM
COLD
COLD
WARM
LOW
LOW
HIGH
HIGH
LOW
LOW
30
Summary

• Potential positive SST-atmosphere feedback over
  the mid-latitude western Pacific
• The atmospheric response appears to be driven
  by both the SST and heat flux forcing the former
  favors a warm-high response, while the latter a
  warm-low response
• The fully coupled model generates, perhaps, the
  most correct atmospheric response, because it
  generates the correct combination of SST and heat
  flux forcing.
• The atmospheric response is best described in the
  coupled system as a part of a least damped
  coupled mode

31
Thank you
32
Hypothesis 1 Flux-SST Forcing Hypothesis

• The SST forcing forces a warm-ridge response,
  while the heat flux forcing forces a warm-low
  response. The combination drives the coupled
  response. (heat flux is the cause !)

SST

• Problem Why the SST and heat flux act
  independently?

33
Hypothesis 2 Eddy-Feedback Forcing Hypothesis

• A warm SST distorts the storm track northward and
  generate a warm-ridge response due to the
  eddy-vorticity forcing. The atmospheric easterly
  wind reduces the surface heat flux loss. The
  stronger the response, the weaker the flux (heat
  flux is the result !)

• Problem Why a stronger SST generates a stronger
  
• warm-ridge response?

34
Hypothesis 3 Coupled Mode Hypothesis

• The atmospheric response is best described in
  the coupled
• system as a part of a least damped coupled mode
• Problem
• What is the role of coupling, on eddy or on flow?