Mechanisms of interannual-decadalvariability in the subtropical-tropical

circulation of the Pacific OceanJ. F. Lübbecke, C. W. Böning and A. Biastoch

IfM-Geomar, Leibniz Institute for Marine Sciences, Kiel, Germany (jluebbecke@ifm-geomar.de)

Abstract:A set of experiments with a 0.5° - version of the global ORCA-LIM model is used toinvestigate the patterns and mechanisms of the interannual-decadal variability in theshallow subtropical-tropical circulation of the Pacific Ocean. The atmospheric forcingbuilds on the bulk formulation and data sets developed by Large and Yeager (2004),based on a combination of NCEP/NCAR reanalysis data and satellite products,following the suggestions for the „Coordinated Ocean-ice Reference Experiments“(CORE) of the Clivar WGOMD. The model captures the salient features of the meancirculation in the tropical Pacific and reproduces the observed decline in the strength ofthe subtropical cells (STCs) and the corresponding warming in the central equatorialPacific over the last decades. The trend in the net equatorward transport across, e.g.,8°S, is governed by a transport change in the interior and partly compensated by a risein the strength of the New Guinea Coastal Undercurrent at the western boundary. Theanticorrelation between interior and western boundary transport changes, found to holdboth at interannual and decadal time scales, can be explained in terms of oscillations inthe horizontal gyre circulation, namely the southward extension of the southern tropicalgyre. Additional experiments, using interannually-varying forcing only for the heat ormomentum fluxes, confirm the role of the remote, wind-driven STC-variability ingenerating equatorial SST-anomalies.

3) Mean Circulation:

Compared to the Levitus climatology the modelSST is too high at the South American coastand features a cold bias of about 1.5 K in theeastern tropics. The latter is typical for oceanmodels and might be due to the fact thatbiology is not taken into consideration and toproblems with lateral mixing.

1) Introduction:The Pacific Subtropical Cells (STCs) aresupposed to affect the equatorial sea surfacetemperature (SST) by changing the amount ofcold water that is brought to the tropics(Kleeman et al., 1999). McPhaden & Zhang(2002) used hydrographic data and observedSST to show that the warming of the tropicalpacific since the 1970s was associated with adecrease in the strength of the STCs. Morerecently (Zhang & McPhaden, 2006) theypointed out that ocean model results showdifferent magnitudes of the slow down of thecells, all being weaker than their observations.Lee & Fukumori (2003) as well as Capotondi etal. (2005) found that the reduction in the STC-strength is confined to the interior ocean andpartly compensated by transports at thewestern boundary in their model studies.

References:

• Capotondi, A., Alexander, M. A. , Deser, C. &McPhaden, M. J., 2005. Anatomy and DecadalEvolution of the Pacific Subtropical-Tropical Cells(J. Climate, 18, 3739-3759)

• Kleeman, R., McCreary, J. P. & Klinger, B. A., 1999.A mechanism for generating ENSO decadalvariability (Geophys. Res. Lett., 26, 1743-1746)

• Large, W. & Yeager, S., 2004. Diurnal to decadalglobal forcing for ocean and sea-ice models: thedata sets and flux climatologies, NCAR TechnicalNote: NCAR/TN-460+SRT (CGD Division of NCAR)

• Lee, T. & Fukomori, I., 2003. Interannual-to-decadalvariations of the tropical-subtropical exchange in thePacific Ocean: Boundary versus interior pycnoclinetransports (J. Climate, 16, 4022-4042)

• Lohmann, K. & Latif. M., 2005. Tropical PacificDecadal Variability and the Subtropical-TropicalCells (J. Climate, 18, 5163-5178)

• McPhaden, M. J. & Zhang, D., 2002. Slowdown ofthe meridional overturning circulation in the upperPacific Ocean (Nature, 415, 603-608)

• Nonaka, M., Xie, S.-P. & McCreary, J., 2002.Decadal variations in the Subtropical Cells andequatorial Pacific SST (Geophys. Res. Lett., 29,20-1– 20-3)

• Zhang, D. & McPhaden, M. J., 2006. Decadalvariability of the shallow Pacific meridionaloverturning circulation and its relation to tropical seasurface temperatures in observations and climatechange models (Ocean Modelling, submitted)

Fig 3:difference in the SST between the last year of a climatological ORCA05 experiment and the Levitus data set

- 1.5 K

2) Model:• ocean - sea ice model ORCA-LIM• ocean model OPA9.0• tripolar grid• horizontal resolution nominal 0.5°, refined at

the equator• vertical 46 layers with variable thickness• atmospheric forcing „CORE“, based on

NCEP/NCAR reanalysis data, corrected byindependent data sets and globally balanced(Large and Yeager, 2004)

• various experiments:

Conclusions:• mean structure of the subtropical-tropical pacific in theORCA model agrees well with results from observationsand other ocean models

• the trends described by McPhaden & Zhang (2002) arereproduced by the ORCA model and are in contrast toother ocean models even larger than observed:

• reduction of STC-strength of total 14 Sv• rise in equatorial SST of about 1.7 K in the Nino3

region• transport variations of the New Guinea CoastalUndercurrent and in the interior are anticorrelated both atinterannual and decadal timescales which can beexplained by changes in the horizontal circulation

• wind-driven EUC- and STC-variability play a role ingenerating SST-anomalies

In the mean meridional overturningstreamfunction one can clearly identify theTropical Cells (TCs) in the range up to 5° N/Sand 100 m depth in both hemispheres as well asthe STCs up to 30°S respectively 25°N and500m depth.

In the meridional section of the mean zonalvelocity at 155°W the most important zonalcurrents can clearly be seen. The EquatorialUndercurrent (EUC) is centered around theequator with a core depth of about 100m and amaximum velocity of 90 cm/s. In general, thisagrees well with observations except that theSubsurface Countercurrents are missing andthat the extension of the EUC is too deep.

Fig 1: different branches of the Subtropical Cells

Fig 2: tripolar ORCA grid

Fig 4: meridional overturning streamfunction for Pacific and Indian Ocean

STC STCTC TC

[m]

EUC

nSECsSEC

NECC

NEC0.9

Fig 5: meridional section at 155°W of the mean zonal velocity

[m]

interannualclimatologicalwind only

climatologicalinterannualheat only

interannualinterannualfully interannual

climatologicalclimatologicalclimatological

wind forcingheat forcingexperiment

interannualclimatologicalwind only

climatologicalinterannualheat only

interannualinterannfully interannual

climatologicalclimatologicalclimatological

wind forcingheat forcingexperiment

interannual

Fig 11: decadal variability of anomalies of the STC Index and SST averaged over 180°W-90°W and 9°S-9°N

4b) SST

As observed, the slowdown of the STCs isassociated with a warming in the easternequatorial pacific.The temperature variations are not only drivenby the thermal forcing but also by windvariations which leads to the question if, inORCA, the wind-driven variability of the STCsand the EUC plays a role in generating SSTanomalies.

Following Nonaka et al. (2002), the SSTvariability in the tropical pacific is at interannualtimescales mainly driven by equatorial winds.Consistent with this exists a good anticorrelationbetween the anomalies of the EUC transportand the SST at 155°W which probably resultsfrom upwelling.

In contrast, at decadal timescales do also off-equatorial winds influence the SST variability atthe equator. Therefore we find an anticorrelationbetween the SST variations and the anomaliesof the STC-Index at decadal timescales. Thisconnection can be explained by the v‘T-mechanism described by Kleeman et al. (1999).

Fig 10: interannual variability of EUC transport and SST at 155°W

correlation r = - 0.89

SSTEUC

fully interannual forced: +1.7 Kheat onlywind only: +1.1 Kheat only + wind only

Fig 9: interannual variability and linear trends of anomalies of the SST averaged over Nino3- region (150°W-90°W, 5°S-5°N)

[K]

a) STCTo investigate the variability of the STCs andTCs, indices for their strength were defined (asin Lohmann & Latif, 2005) by taking themaximum of the meridional overturning minusthe minimum in a certain range. Both indicesdecreased with a reduction of 28 Sv for theTCs and 14 Sv for the STCs.

Fig 8: barotropic streamfunction averaged over (a) 1965 to 1976 respectively (b) 1977 to 1997; red line indicates 8°S

The total decrease of the STC is governed by atransport change in the interior and partlycompensated by a rise in the strength of thewestern boundary current. Because figure 7shows the situation only for the southern cell,the reduction in the interior STC transport isfound to be larger than observed.The variations of the western boundary currentand the interior transport are generallyanticorrelated to each other as described by Lee& Fukumori (2003).

The anticorrelation can be explained by changesin the horizontal circulation, namely a southwardextension of the tropical gyre, which leads to anintensification at 8°S. A stronger tropical gyrestrengthens the western boundary current whileit counteracts the interior exchange window. Theextension itself is a result of a modified windforcing (Lee & Fukumori, 2003) and theadjustment of the ocean circulation isaccomplished through the propagation ofbaroclinic Rossby waves (Capotondi et al.,2005).

Fig 6: timeseries of the indices for STCs and TCs, low-pass filtered with 23 points

linear trends:STC: -14 SvTC: -28 Sv

[Sv]

Fig 7: net equatorward transport across 8°S in the STC – density range (22.7<σ<26.8)

total

interior

wbc[Sv]

linear trends: -7 Sv in total+7 Sv wbc-14 Sv interior

4) Interannual to decadal variability: