Pacific vs. Indian Ocean warming: How does it matter for global and regional climate change?

1
Pacific vs. Indian Ocean warmingHow does it
matter for global and regional climate change?

• Joseph J. Barsugli
• Sang-Ik Shin
• Prashant D. Sardeshmukh
• NOAA-CIRES Climate Diagnostics Center
• Boulder, Colorado

NOAA Photo Library
2
Q How does it matter for global (DJF) climate
change?
A Opposing Temperature and Precipitation
Sensitivity to Tropical SSTs across 110E
3
Q How does it matter for regional (NH
Wintertime) climate change?
A Opposing PNA and NAO Sensitivity to Tropical
SSTs across 110-120E, with PNA emphasizing the
central and eastern Pacific more.
4

• Method and Context (How did we get these plots
  and why do we believe them?)
• Understanding Opposing T, p sensitivity
• Seasonal cycle in T, p sensitivity
• PNA and NAO sensitivity with application to 50
  year trend.

5
SST Patch Experiments with NCAR CCM3.10
1 K 0.5K

• 43 SST anomaly patches total. 2K maximum
  anomaly
• Indo-Pacific
• 32 ensemble members (16 Warm, 16 Cold)
• 18 month runs (2 winters, 1 summer)
• Atlantic
• 40 ensemble members (20 warm, 20 Cold)
• 25 month runs
• Climatology
• 100 year fixed climo-SST run

6
Comparison of the present study to Barsugli and
Sardeshmukh, 2002
Atmosphere GCM

• NCAR CCM3.10 w/CCM3.6 physics
• T42, 18 level resolution
• Default parameter values for current climate
• Smoothed Topography (as in CAM2)

• CCM3 vs. NCEP MRF9
• Larger patch area in Indo-Pacific.
• 2K vs. 1.5K peak SST anomaly.
• Seasonal cycle vs. perpetual January
• All Tropical oceans vs. Indo-pacific only.

7
Green Function approximation

• Think of this as an empirical linearization
  about a given climate.
• For a grid of T, or for a set of patches this
  becomes
• Note the Area Factor. This means that the
  grid-independent Green Function (and sensitivity)
  has units of
• Indo-Pacific Patches 12 SSTU
  Atlantic Patches 7.5 SSTU

8
Barsugli and Sardeshmukh, 2002
9
Simmons, Wallace, and Branstator, 1983

• Barotropic vorticity forcing
• nodal line for Pacific target? This is similar
  to patterns in Branstator et al, 1985 from
  explicit Green function approach.
• Hmmm, what about the forcing of North Atlantic.
• Newman and Sardeshmukh, 1998. Strong seasonal
  dependence of sensitivity.

10
Ting and Sardeshmukh, 1993

• Linear Baroclinic Model. Deep heating.
• Remote response switches sign across 120E.
• Little remote response near nodal line at 120E.
• Local (rotational) response moves with forcing
• Heating is used as forcing, not SST.
• Nodal line not found when GFDL model basic state
  is used.

11

• Hoerling et al. (2004). Indian ocean SSTs force
  50 year trend in North Atlantic NAO-like
  pattern. Found in 3 models.
• Branstator(2004) Circumglobal Waveguide Pattern
  responds more to forcing at the date line than to
  forcing at 150 W. The latter forcing results in
  a more isolated PNA pattern.

12
Temperature and Precipitation Sensitivity
13
(No Transcript)
14
Composite Surface Air Temperature
Indian Ocean
West Pacific Ocean
15
Composite Precipitation
Indian Ocean
West Pacific Ocean
16
Indian Ocean Patch Composite
Surface Latent Heat Flux
U-Wind Stress
V-Wind Stress
500 hPa Omega
17
W. Pacific Ocean Patch Composite
Surface Latent Heat Flux
U-Wind Stress
V-Wind Stress
500 hPa Omega
18
Precipitation Sensitivity Seasonal Cycle
DJF
MAM
JJA
SON
ANNUAL
19
Surface Temperature
20
850 hPa Temperature
21
Annual Mean Temperature at different levels
22
NAO, PNA and Trends
23
NAO and PNA sensitivity
- NAO
PNA
PNA MRF9
24
PNA in detail
25
1950-1999 Trend in 500 hPa Z NCEP
Reanalysis
Atlantic Sector
Pacific Sector
26
Z500 trend from CCM3
TOGA and GOGA
Fixed SST Trend Pattern
27
Projection of Z500 trend onto each patch
experiment
TOTAL TREND
PAC
ATL
28
Pattern Correlation
29
Conclusions

• There is opposing sensitivity to SST anomalies
  across longitude 110E for global mean
  temperature and precipitation. E.g. warm SST
  anomalies in most of the Tropical Indian Ocean
  lead to a global mean cooling of the surface (
  mainly land), cooling at 850 mb, and a reduction
  in global mean precipitation.
• The opposing temperature sensitivity is confined
  to the DJF season and is largest in the northern
  continents. The opposing sensitivity for
  precipitation is evident all year, and there is
  an additional area of negative sensitivity in the
  Eastern Pacific, south of the Equator.
• There is opposing sensitivity to SST anomalies
  across the same nodal line for the NAO and PNA
  pattern. Compared to the NAO, the PNA shows more
  sensitivity to SSTs at and east of the dateline.
  
• Therefore, the relative warming of the Indian and
  West Pacific warm pools will have a large impact
  on both global and (Northern Hemisphere) regional
  response to Tropical SST changes in DJF.
  Prediction (or past and paleo- reconstruction)
  of this broad spatial pattern of SST change is
  essential to get the global and regional picture
  correct.

30
Conclusions

• Because the nodal line was seen in BS2002 and
  earlier, more idealized dynamical studies, we
  believe it to be a robust structure in the
  atmosphere of dynamical origin. The temperature
  sensitivity follows from the circulation
  anomalies. However, the origin of the coincident
  nodal line in Tropical precipitation remains
  uncertain.
• The Tropical Atlantic SSTs are generally more
  influential in JJA for the global mean
  quantities.
• The full picture of Tropical-Extratropical
  interaction is a) more complicated than just
  ENSO, but not that much more complicated 2-3
  regions do pretty well to capture the changes
  that have a big global impact.
• Prashant will elaborate on the big picture at
  tomorrows talk.