Michael J. McPhaden, NOAA/PMEL Dongxiao Zhang, University of Washington and NOAA/PMEL Circulation...

Michael J. McPhaden, NOAA/PMEL

Dongxiao Zhang, University of Washington and NOAA/PMEL

Circulation Changes Linked to ENSO-like Pacific Decadal Variability

Workshop on Low Frequency Modulation of ENSO

23 September 2003

Toulouse, France

Purpose

1) Review observations indicating a slowdown in the shallow meridional overturning circulation in the tropical Pacific from the 1970s to the 1990s coincident with the development of warm PDO conditions (McPhaden & Zhang, Nature, 2002).

2) Update these results for the late 1990s to the present.

Circulation Changes Linked to ENSO-like Pacific Decadal Variability

S. Hare, personal comm.

Relevance of PDO

Affects the climate of North America (Latif & Barnett, 1994; Cayan et al, 2001) Affects marine ecosystems of the Pacific (Mantua et al, 1997; Hare and Mantua, 2000; Chavez et al, 2003)

May affect NAO and Atlantic deep meridional overturning circulation (Latif 2001; Hoerling et al, 2001)

Linked to decadal modulation of ENSO (Trenberth and Hurrell, 1997; Latif et al, 1997)

Relevance of PDO

Affects the climate of North America (Latif & Barnett, 1994; Cayan et al, 2001) Affects marine ecosystems of the Pacific (Mantua et al, 1997; Hare and Mantua, 2000; Chavez et al, 2003)

May affect NAO and Atlantic deep meridional overturning circulation (Latif 2001; Hoerling et al, 2001)

Linked to decadal modulation of ENSO (Trenberth and Hurrell, 1997; Latif et al, 1997)

Theories and Hypotheses*

Mid-latitude ocean-atmosphere interactions (Latif & Barnett, 1994; Yuleava et al, 2001) Tropical ocean-atmosphere interactions (Trenberth and Hurrell, 1994; Graham, 1994) Tropical-extratropical ocean-atmosphere interactions (Gu and Philander, 1997; Kleeman et al, 1999)

Stochastically forced by atmospheric noise (Burgers, 1999; Barnett et al, 1999)

* Reviewed by Miller and Schneider, 2000, Prog. Oceanogr.

Kleeman et al (1999) Hypothesis for Pacific Decadal Climate Variability:

Decadal time scale tropical Pacific temperature anomalies are determined by the rate at which the

subtropical cells transport thermocline water towards the equator (V’T) rather than by the

transport of anomalously warm or cold thermocline water by the mean circulation (VT’).

Ocean Circulation Schematic

Mean Circulation in Pycnocline

9°N

9°S

14 Sv

7 Sv

(Integrated over 22.5-26.5 kg m-3)

Changes from 1970s to 1990s

Interior Ocean Pycnocline Transport Changes(22.5-26.5 kg m-3)

McPhaden & Zhang, 2002

SST anomaly (9°N-9°S, 90°W-180°)

Surface Layer Divergence

*Ekman transports computed from NCEP, ECMWF, COADS, FSU wind stresses; surface transport adjusted for geostrophic flow

*

positive phase

Change in 1998?

0.84

-0.36

“The Perfect Ocean for Drought” (Hoerling & Kumar, Science, 2003)

“…the modeling results offer compelling evidence that the widespread drought was strongly determined by the tropical oceans.”

Reynolds SST; ERS & Quikscat wind stress

Reynolds SST; ERS & Quikscat wind stress

Reynolds SST; ERS & Quikscat wind stress, TOPEX/Poseidon & Jason sea level

Potential Vorticity (= 25 kg m-3) CTD Casts to 900 m

July 92-June 98

July 98-June 03

11,585

6,729

Interior Ocean Pycnocline Transport Changes

Pacific Ocean Circulation Changes9°N to 9°S

13.0

49.7

36.7

42.4

28.1

57.9

29.8

51.9

0

10

20

30

40

50

60

70

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1992-1998

1998-2003

Surface layer divergence is Ekman divergence (based on ERS & Quikscat wind stress) reduced by geostrophic convergence in the surface layer.

Pacific Ocean Circulation Changes9°N to 9°S

13.0

49.7

36.7

42.4

28.1

57.9

29.8

51.9

0

10

20

30

40

50

60

70

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1992-1998

1998-2003

Western boundary current transport convergence computed assuming mass conservation (surface layer divergence minus pycnocline convergence) and Indonesian Throughflow decadal variations small.

Pacific Ocean Circulation Changes9°N to 9°S

13.0

49.7

36.7

42.4

28.1

57.9

29.8

51.9

0

10

20

30

40

50

60

70

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1992-1998

1998-2003

Equatorial upwelling computed assuming 20% of western boundary current transports in the surface layer, 80% in the pycnocline.

Western Boundary Current Compensation

Heat and mass fluxes into and out of the interior tropical Pacific Ocean are partially compensated by flows in the western boundary currents on seasonal-to-interannual time scales (Cane and Sarachik, 1979; Springer et al, 1990).

Pacific Ocean Circulation Changes9°N to 9°S

27.0

51.7

24.7

46.8

28.1

57.9

29.8

51.9

0

10

20

30

40

50

60

70

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1970-1977

1998-2003

Differences between the 1970-1977 and 1998-2003 are smaller than differences between either of these periods and early-mid 1990s; differences between 1970-77 and 1998-2003 may also not be significant given computational uncertainties.

CMAP rainfall; ERS & Quikscat wind velocity

CMAP rainfall; ERS & Quikscat wind velocity

1) The shallow meridional overturning circulation in the Pacific has accelerated since the late 1990s in concert with a cold phase shift in the PDO.

Conclusions

2) Pycnocline, surface layer, and equatorial upwelling transports are presently comparable to those prior to the mid-1970s “regime shift.”

3) As on seasonal-to-interannual time scales, changes in interior ocean circulation on decadal time scales are partially compensated for by opposing changes in western boundary current transports (~1/3)

4) Sudden reversal of tropical Pacific warming and associated circulation changes suggest that greenhouse gas forcing effects on the 1970s to 1990s warming trend were of secondary importance.

Conclusions

5) Structural and dynamical similarities between ENSO and the tropical manifestations of the PDO make it very difficult to deconvolve cause and effect between the PDO and decadal modulation of ENSO.

Interannual ENSO cycle and tropical Pacific manifestations of the PDO exhibit analogous patterns of variability and share similar dynamics, though with different time scales.

PDO cold phase

PDO warm phase

Ocean Circulation Schematic

Mean Circulation in Pycnocline

Data Sets

Wind data sets:

COADS ship/buoy, 1945-93

NCEP reanalysis, 1958-99

ECMWF reanalysis and operational analyses, 1979-99

FSU ship/buoy, 1961-99

Meridional Structure

Mean Circulation in Pycnocline

Ekman Transport

9°N

9°S

Pacific Ocean Circulation Changes9°N to 9°S

27.0

51.7

24.7

46.8

14.0

40.7

26.7

35.4

0.0

10.0

20.0

30.0

40.0

50.0

60.0

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1970-1977

1990-1999

Western boundary current transport convergence computed assuming mass conservation (surface layer divergence minus pycnocline convergence).

Pacific Ocean Circulation Changes9°N to 9°S

27.0

51.7

24.7

46.8

14.0

40.7

26.7

35.4

0.0

10.0

20.0

30.0

40.0

50.0

60.0

PycnoclineConvergence

Surface LayerDivergence

WesternBoundaryCurrent

EquatorialUpwelling

Volume Transport (Sv)

1970-1977

1990-1999

Equatorial upwelling computed assuming 20% of western boundary current transports in the surface layer, 80% in the pycnocline.