ENSOforced variability of the Pacific Decadal Oscillation

1
ENSO-forced variability of the Pacific Decadal
Oscillation

• Matt Newman
• NOAA-CIRES CDC

See Newman, Compo, and Alexander, Dec. 2003, J.
Climate.
2
Is the PDO independent of ENSO?
3
From Alexander et al 2002
4
Re-emergence
Depth (m)
From Alexander et al 2002
5
Null hypothesis

• PDO as an AR1 (red) process
• ENSO forcing

6
Data

• Hadley Ice Sea Surface Temperature Analysis
  (1900-1999) and Reynolds SST (2000-2001)
• Monthly mean SST anomalies determined from
  1950-2001 annually-varying climatology
• Same results (not shown) for 1950-2001 period
  with Reynolds Reconstructed SST

7
Indices 1900-2001

• PDO (Pacific Decadal Oscillation)
• Projection of North Pacific SST upon leading EOF
  of monthly North Pacific (north of 20N) SST for
  the years 1950-2001 (Mantua et al 1997)
• ENSO (Leading tropical Pacific PC)
• Projection of tropical Pacific SST upon leading
  EOF of monthly tropical Pacific (20N-20S,
  120E-60W) SST for the years 1950-2001
• NPI (North Pacific Index)
• Area-weighted SLP in region 30-65 N, 160E-140W
  (Trenberth and Hurrell 1994)
• All indices smoothed with 3-month or 12-month
  running means 102-yr linear trend removed

8
Indices
(note not detrended here, but centered)
9
Annual cycle of lead-lag correlation (1950-2001)
10
(No Transcript)
11
Wavelet analysis (1900-2001)
12
PDO and ENSO are coherent on both interannual and
decadal periods
13
variance of decadal averages variance of
seasonal averages PDO 0.54 ENSO 0.32 NPI 0.30
14
A null hypothesis
where
July-June mean PDO index
July-June mean ENSO index
white noise
applied to annual mean anomalies
15
Forecast PDO vs. observed PDO
r 0.74
Forecasts are cross-validated
Skill is significantly better (95 level) than
AR1 model
16
Power spectra (annual means)
PDO response to ENSO forcing is redder than ENSO
forcing
17
ENSO-PDO relationshipA good test for coupled
GCMs
18
Comparison of observed power spectra (gray
shading) with CMIP2power spectraannual means
19
Null hypothesis applied to CMIP2 output
20
Summary Slide

• Null hypothesis of the PDO
• North Pacific SST integrates effects of ENSO via
  the atmospheric bridge
• Re-emergence brings back ENSO-induced anomalies
  in succeeding winters (no summer/fall PDO)
• PDO should be redder than ENSO, e.g.

21
North Pacific SST anomaly should be greater than
Tropical Pacific SST anomaly on decadal
timescales
22
Penultimate Slide

• Consequences for analysis
• Dont stratify ENSO into high/low PDO years?
• Summertime PDO just ENSO?
• Spring PDO more than ENSO?
• What other climate integrators might redden
  ENSO?
• Tree rings
• Drought (soil moisture anomalies)

23
Last Slide

• What part of predictable North Pacific decadal
  variability is independent of the Tropics?
• Data is limited
• Coupled GCMS must reproduce not only ENSO power
  spectrum
• but also seasonal cycle of a and b

24
PDO forecast

• Continued positive next year
• (assuming that we do not have a cold event in
  the Tropical Pacific)
• For more, see Newman, Compo, and Alexander, Dec.
  2003, J. Climate, 16, 3853-3857.