«and Wilhelmus P. M. De Ruijter Institute for Marine and Atmospheric Research Utrecht Utrecht University Utrecht, the Netherlands and Andreas Sterl ...»
Response of the Atlantic overturning circulation
to South Atlantic sources of buoyancy
Wilbert Weijer and Wilhelmus P. M. De Ruijter
Institute for Marine and Atmospheric Research Utrecht
Utrecht, the Netherlands
Andreas Sterl and Sybren S. Drijfhout
Royal Netherlands Meteorological Institute
De Bilt, the Netherlands
Submitted to: Global and Planetary Change
Revised version of February 2, 2001
Wilbert Weijer Institute for Marine and Atmospheric Research Utrecht Department of Physics and Astronomy Utrecht University Princetonplein 5, 3584 CC Utrecht The Netherlands Phone: ++31-30-2537759; Fax: ++31-30-2543163 Email: W.Weijer@phys.uu.nl Abstract The heat and salt input from the Indian to Atlantic Oceans by Agulhas Leakage is found to inﬂuence the Atlantic overturning circulation in a low-resolution Ocean General Circulation Model. The model used is the Hamburg Large-Scale Geostrophic (LSG) model, which is forced by mixed boundary conditions. Agulhas Leakage is parameterized by sources of heat and salt in the upper South Atlantic Ocean, that extend well into the intermediate layers.
It is shown that the model’s overturning circulation is sensitive to the applied sources of heat and salt. The response of the overturning strength to changes in the source amplitudes is mainly linear, interrupted once by a stepwise change.
The South Atlantic buoyancy sources inﬂuence the Atlantic overturning strength by modifying the basin-scale meridional density and pressure gradients. The nonlinear, stepwise response is caused by abrupt changes in the convective activity in the northern North Atlantic.
Two additional experiments illustrate the adjustment of the overturning circulation upon sudden introduction of heat and salt sources in the South Atlantic.
The North Atlantic overturning circulation responds within a few years after the sources are switched on. This is the time it takes for barotropic and baroclinic Kelvin waves to reach the northern North Atlantic. The advection of the anomalies takes 3 decades to reach the northern North Atlantic.
The model results give support to the hypothesis that the re-opening of the Agulhas Gap at the end of the last ice-age, as indicated by palaeoclimatological data, may have stimulated the coincident strengthening of the Atlantic overturning circulation.
• thermohaline circulation
• ocean circulation
• South Atlantic 1 Introduction In the subpolar regions, the surface waters of the Atlantic are by far the saltiest in the World Ocean (e.g., Levitus 1982), and contrast strongly with surface waters in, for instance, the North Paciﬁc. This contrast is thought to be one of the factors responsible for deep water being formed mainly in the North Atlantic, and not in the North Paciﬁc (Warren 1983). The high salinities of the Atlantic thermocline waters are in general ascribed to the excess evaporation and associated freshwater export from the Atlantic basin, that is estimated at 0.7 Sv north of 35◦ S (Baumgartner and Reichel 1975).
Gordon (1985, 1986) was the ﬁrst to recognize that the exchange of water between the South Indian and Atlantic Oceans might play a role in maintaining the high Atlantic surface salinities as well. This exchange, commonly known as Agulhas Leakage (Lutjeharms 1996; De Ruijter et al. 1999), is mainly accomplished by the shedding of large rings from the Agulhas Current in the so-called Agulhas Retroﬂection area.
Filled with warm and salty Indian Ocean water, these rings subsequently drift into the South Atlantic, where they dissipate. The water entering the Atlantic through this leakage is saltier than the surrounding South Atlantic waters (Gordon et al. 1987). It is derived mainly from the evaporative subtropical Indian Ocean (Indian Ocean Central Water, Gordon et al. 1992), and it is subjected to strong evaporative activity in the Agulhas Retroﬂection area (Gordon et al. 1987; Van Ballegooyen et al. 1994). On the basis of this observation, Gordon et al. (1992) suggested that Agulhas Leakage could have dynamical impact on the meridional overturning circulation as well: the salt input saliniﬁes the Atlantic surface waters and would precondition them for the formation of NADW. They even speculated: “if the Indian Ocean salt input were severed, might the NADW thermohaline cell run down?” [Figure 1 about here.] The answer to this question may have implications for the thermohaline circulation during the Pleistocene. De Ruijter (1982) has shown that the presence of Agulhas Leakage critically depends on the position of the Subtropical Convergence Zone (STCZ), that may be considered as the southern boundary of the wind-driven gyres of the South Indian and Atlantic Oceans. Compared to the present-day wind ﬁeld, a northward shift of only a few degrees would suﬃce to eﬀectively shut oﬀ the connection between the Subtropical Gyres of the South Indian and South Atlantic Oceans.
A study of planktonic foraminifera assemblages in several sediment cores from the South Indian Ocean (Howard and Prell 1992) indicates that the positions of the major Southern Ocean fronts have shifted considerably during the last 500 kyr. These shifts mainly reﬂect the major glacial-interglacial cycles. In particular, the STCZ was shifted several degrees northward during the glacial periods with respect to its present position at about 45◦ S (Fig. 1). Analysis of calcareous plankton assemblages from a core recovered from the Cape Basin (Flores et al. 1999) indicates that during glacial stages 2-4 and 6 the inﬂuence of the Agulhas Current on the basin was strongly reduced. This is consistent with a more northerly position of the STCZ, and an associated reduction of the leakage.
Berger and Wefer (1996) pointed out that the establishment of a ﬁrm Atlantic overturning circulation at the end of the last ice-age coincided with the reappearance of the foraminifera species Globorotalia menardii in the tropical Atlantic. This species had been extinct in the glacial Atlantic, but not in the Indian Ocean. Its reappearance can only be explained by reseeding of the Atlantic from the Indian Ocean, possibly in response to a reopening of the Agulhas connection. Bearing in mind the hypothesis of Gordon et al. (1992), Berger and Wefer (1996) speculated whether the reopening of Agulhas Leakage may have played a role in the sequence of events leading to the recovery of the overturning circulation.
Cai and Greatbatch (1995) were the ﬁrst to address the hypothesis of Gordon et al.
(1992) by comparing two models of the global ocean circulation; in one model Agulhas Leakage was present, whereas in the other it was inhibited by extending the African continent several degrees southward. They concluded that the strength of the Atlantic overturning circulation is unaﬀected by the presence or absence of Agulhas Leakage.
However, Rahmstorf et al. (1996) pointed at the unrealistically weak thermal coupling between the ocean and atmosphere that was implied by the so-called zero heat capacity atmosphere model used by Cai and Greatbatch (1995). This severely hindered the development of density anomalies, and hence the pressure ﬁeld was not aﬀected. Furthermore, their approach did not allow for systematically varying the Agulhas exchange, and the robustness of the response was neither questioned nor tested.
Weijer et al. (1999, 2001) showed that under mixed boundary conditions (i.e., prescribed freshwater ﬂux and relaxation of sea-surface temperature towards a prescribed proﬁle) inter-ocean ﬂuxes of heat and salt may inﬂuence the overturning circulation considerably. This type of surface forcing allows for the development of density anomalies, since thermal anomalies are rapidly damped while saline anomalies are unaﬀected.
However, their model was highly simpliﬁed, and represented the overturning circulation in a 2D meridional-depth plane of the Atlantic Ocean. Consequently, it lacked realistic features that probably aﬀect the sensitivity of the overturning circulation in the real ocean, like basin geometry and bathymetry, the wind-driven circulation and rotational eﬀects. Furthermore, only steady states were calculated, so that the adjustment of the overturning circulation upon sudden changes in the inter-ocean exchange could not be studied.
In the present study, the question of the impact of Agulhas Leakage on the overturning circulation is addressed using the low-resolution Hamburg LSG model. Due to its coarse spatial resolution and high viscosity, the model does not resolve Agulhas Leakage under realistic wind stress forcing (Drijfhout et al. 1996). Therefore, the Agulhas heat and salt exchanges are parameterized by heat and salt sources in the South Atlantic (and equally large heat and salt sinks in the South West Indian Ocean, from which Agulhas Leakage originates). In this way the thermohaline impact of Agulhas Leakage can be studied systematically, without explicitly aﬀecting the momentum balance. It will be shown that in this model the Agulhas heat and salt sources aﬀect the overturning strength by changing the large-scale density and pressure distributions (as suggested by the studies of Hughes and Weaver (1994) and Rahmstorf (1996)), rather than by changing the convective activity (as suggested by Gordon et al. (1992)). Additionally, the adjustment of the overturning circulation is studied to changes in the Agulhas source amplitudes. This will shed some light on the mechanisms that generate the response of the overturning circulation, and on their corresponding time scales.
The model is introduced in section 2. The spin-up procedure is described, and an evaluation is made of the circulation that is used as starting point for the experiments.
The section concludes with a description of the source/sink parameterization that is used to model the Agulhas input of heat and salt. In section 3 the results of the main experiment are presented and analyzed, concerning an integration over 10 kyr. In section 4 the results of two adjustment experiments are analyzed. Section 5 discusses the implications of the results, and evaluates the Agulhas Leakage parameterization.
2.1 The Hamburg LSG model The Hamburg Large-Scale Geostrophic (LSG) model has been developed for climate studies, and it was described in detail in Maier-Reimer and Mikolajewicz (1992) and Maier-Reimer et al. (1993). In this model the fast modes have been ﬁltered out by integrating the complete set of primitive equations with a semi-implicit time-stepping method, and by neglecting non-linear advection of momentum. This allows for a time step of 30 days. The model has a free-surface formulation, with a sea-surface height (ζ) that is prognostically determined. The ocean is coupled to a simple thermodynamic sea-ice model, and is forced by the wind-stress climatology of Hellerman and Rosenstein (1983). The model is discretized on an Arakawa E-grid (Arakawa and Lamb 1977). It has an eﬀective horizontal resolution of 3.5◦ × 3.5◦, and 11 vertical levels, centered at depths of 25, 75, 150, 250, 450, 700, 1000, 2000, 3000, 4000 and 5000 m. For the momentum equations, the standard hydrostatic and Boussinesq approximations are applied, and vertical friction is neglected.
The fully non-linear unesco (1981) equation of state is used. Furthermore, a convective adjustment scheme is applied that interchanges unstably stratiﬁed pairs of layers in one downward sweep. This removes most, though not all, of the instabilities. Bryden’s (1973) polynomial expression is used to convert the potential temperature θ to the in-situ temperature T (Maier-Reimer et al. (1993) use a linearized equation). For reasons given below, a higher value of the viscosity (4 · 105 m2 s−1 ) is used in this study than the 5 · 104 m2 s−1 used by Maier-Reimer et al. (1993).
2.2 The source distribution
Heat and salt exchange by Agulhas Leakage takes place by the inﬂow of Indian Ocean water into the south-east Atlantic, and a compensating return transport leaving the Atlantic. The inﬂowing Indian Ocean water is characterized by higher temperatures and salinities than the surrounding South Atlantic thermocline waters. Based on this diﬀerence in water mass properties, Van Ballegooyen et al. (1994) estimated the heat and salt exchange by an average of 6 Agulhas rings per year at about 0.045 PW and 2.52 Gg/s. This is probably an underestimation: Agulhas Leakage is not only sustained by the shedding of large rings, but also by Agulhas Current water leaking directly into the Atlantic (Gordon et al. 1987), partly in the form of ﬁlaments (Lutjeharms and Cooper 1996). Moreover, the exact amount of heat and salt that is exchanged can only be determined when the temperature and salinity is known of the water that is exported to compensate for the Agulhas water input. If, for instance, a large part of the Agulhas inﬂow is compensated by outﬂow of NADW, then the corresponding heat ﬂux may be up to an order of magnitude larger (Gordon 1985).
In this model, Agulhas heat and salt input is parameterized by sources of heat and salt in the South Atlantic Ocean (denoted by qAtlantic ). To close the heat and salt budgets, equal amounts of heat and salt are extracted from the south-west Indian Ocean, where the leakage originates (qIndian ). The strength of the sources are controlled by a dimensionless amplitude σ, that relates the volume integrals of the sources to the
original Van Ballegooyen et al. (1994) values:
where the brackets denote volume integration. A value of σ = 1 thus represents the original Van Ballegooyen et al. (1994) heat and salt exchange of 0.045 PW and 2.52 Gg/s, whereas σ = 10 would be closer to Gordon’s (1985) heat ﬂux estimate.