Abstract

Using a set of models, including one with a resolution of ¼°, several aspects of the simulated seasonal currents in the deep ocean are considered. It is shown that over vast areas of the deep interior, particularly in the Indian Ocean, annual-mean circulation represents a small residual of much stronger seasonal flows. In many places the seasonal horizontal velocities are of the order of 10−2 m s−1, reaching locally to 10−1 m s−1; the corresponding vertical velocities are of the order of 10−5 m s−1. An idealized geometry model is employed to confirm the notion that much of this seasonal variability in the deep-ocean circulation can be attributed to the annual cycle of wind stress, combined with the significant increase in the vertical trapping depth for basin-scale seasonal forcing. It is suggested that, at least on seasonal time scales, the so-called bottom pressure torque can be an important term in the depth-integrated vorticity balance. An interaction of these relatively strong flows (of nontidal origin) with bottom topography may contribute to diapycnal mixing in the deep ocean in a manner similar to that proposed recently for the Southern Ocean. In addition, it is found that under a plausible climate change scenario, the amplitude of the mean annual cycle of wind stress may change. Among the regions where such changes are most pronounced is that in the extratropical North Pacific. It is shown that the data on surface wind stress can be effectively used to identify the seasons with the largest changes in the deep-reaching overturning cells. Finally, unlike what might be expected from the earlier theories, the annual-mean circulation simulated by the model with ¼° resolution has the deep interior flows that tend to group into jetlike structures, often having a predominant equatorward rather than poleward direction.

1. Introduction

In the first attempts to model the large-scale circulation of the deep ocean it was essentially assumed that the interior flows have a simple structure and are weak (Stommel and Arons 1960). Since then, the notion that the deep interior may be dominated by weak circulation has been often repeated. Recent measurements, however, have revealed a more complex picture with relatively strong zonal currents (Hogg and Owens 1999). In addition, there is evidence suggesting that deep flows may vary strongly on seasonal time scales, perhaps mostly resulting from the annual cycle of wind stress. Some evidence that this might be the case comes from observations. In particular, in several places relatively strong deep currents have been observed that seem to have been forced by time-varying winds (Koblinsky and Niiler 1982; Niiler and Koblinsky 1985; Brink 1989; Niiler et al. 1993). Modeling results suggest that this might be a widespread phenomenon. In particular, it has been shown that the directly wind-driven overturning can penetrate to large depths on seasonal time scales, resulting in a large-amplitude (of the order of 5 PW) annual cycle of heat transport in the ocean (Bryan 1982; Boning and Herrmann 1994; Nakano et al. 1999; Jayne and Marotzke 2001). Such a deep penetration of the seasonal overturning has been related to the idea that the so-called vertical trapping scale of the ocean’s response to perturbations (Willebrand et al. 1980; see also next section) is a function of forcing scale and frequency (Jayne and Marotzke 2001). For a basin-scale seasonal forcing, this vertical trapping is comparable to the depth of the ocean.

While models suggest that the amplitude of the deep-reaching seasonal overturning can be as large as 50 Sv (1 Sv ≡ 106 m3 s−1), the associated horizontal velocities in the deep ocean have been stated to be only of the order of 10−3 m s−1 (Jayne and Marotzke 2001). The latter estimate, however, applies only to zonally averaged meridional velocities. One of the purposes here is to illustrate that locally, seasonal currents could be much stronger. In particular, using a high-resolution near-global model we show that over vast areas of the deep-oceanic interior, annual-mean flows may represent small residuals of quite strong seasonal currents with velocities reaching locally 10−1 m s−1. An idealized basin model is employed to illustrate that much of this seasonal variability in the abyss is due to seasonal variability in the wind stress.

The idea that on seasonal time scales vast areas of the abyssal ocean interior could be dominated by flows with speeds exceeding 10−2 m s−1 would have important implications. First, it can be demonstrated, using models or even simple scaling arguments, that in such a case the topographically induced vertical velocity (arising due to cross-slope flows) should be at least of the same order as the near-surface Ekman pumping velocity; hence, the so-called Sverdrup balance is unlikely to hold on seasonal time scales. Second, large-scale and mesoscale flows, with the strength ultimately depending on the strength of wind field and in particular their interaction with topography, are considered among the important sources of internal wave energy-sustaining small-scale turbulence (e.g., Wunsch and Ferrari 2004). However, the wind field operates at the surface, whereas, for example, tidal forcing is introduced at great depth and is thus thought to be efficiently placed for deep mixing (Wunsch and Ferrari 2004). We shall illustrate, however, that on seasonal time scales strong and directly wind-driven currents can penetrate to great depths over vast areas of the oceanic interior, thereby opening the possibility for the deep mixing to be sustained in part by interaction of these flows with topography.

Another related topic touched here concerns the possibility for the mean annual cycle of wind stress to change in response to plausible changes in the climate. In particular, using a climate model forced by a representative radiative forcing scenario, we evaluate the changes in the annual cycle of wind stress projected by the end of the twenty-first century. It is found that the mean annual cycle of wind stress does respond to the changing climate conditions, so that there might be expected changes in the seasonal flows in the deep ocean.

2. Idealized basin and global coarse-resolution models

For many purposes, particularly outside of the tropics, the essential dynamics of the seasonal variability in the ocean driven by seasonal variability in the wind stress curl can be represented by simple models (e.g., White 1978; Qiu 2002). In this section, in order to illustrate a basin-wide penetration of strong seasonal flows to the abyss, we use general circulation models, albeit starting with a model having idealized basin geometry. The idealized geometry model is the Geophysical Fluid Dynamics Laboratory (GFDL) Modular Ocean Model (MOM), version 2, coupled to a simple energy–moisture balance atmosphere model and sea ice model. The model components and the coupling scheme are described in Weaver et al. (2001). Here, the model basin is configured to have a simple geometry. The ocean spans about 70° latitude and about 60° longitude both north and south of the equator. For the illustrative purposes here, the horizontal resolution is 1.8° latitude × 3.6° longitude. Mesoscale eddies are parameterized using the Gent and Mc Williams (1990) mixing scheme. There are 50 vertical levels in the ocean model that vary smoothly in thickness from 20 m at the surface to about 200 m at the deepest levels. Vertical diffusivity has a highly idealized shape: it is set to 10−4 m2 s−1 in the uppermost two levels (upper 43 m), whereas in the rest of the model domain it is set to 0.1 × 10−4 m2 s−1. Vertical viscosity is 50 × 10−4 m2 s−1. The model domain has a circumpolar gap between about 45° and 55°S. The ocean bottom is flat with a depth of about 5 km, except for the ridge within the circumpolar gap that has a depth of about 2.6 km. Both components of wind stress are prescribed using the 40-yr monthly mean National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) climatology, averaged zonally (Fig. 1).

Fig. 1.

Mean meridional profiles of zonally averaged zonal wind stress over the ocean (derived from the 40-yr NCEP–NCAR reanalysis).

Fig. 1.

Mean meridional profiles of zonally averaged zonal wind stress over the ocean (derived from the 40-yr NCEP–NCAR reanalysis).

At the steady state, the simulated annual-mean horizontal streamfunction has a familiar pattern (Fig. 2a). At the latitudes blocked by land, the interior gyres are closed in the west by the boundary currents, the structure of which is largely set by the model’s resolution (i.e., by the horizontal viscosity). The annual-mean, depth-integrated circulation is mainly composed of the flows in the upper ocean, particularly in the low-latitude regions where stratification is strong. Such a confinement to the upper ocean is expected once a sufficient time is allowed for the first baroclinic Rossby mode to propagate (Anderson and Gill 1975). In addition, because of the imposed small value for the vertical diffusivity, the annually averaged ocean currents below about 2 km, including at the western boundary, are essentially absent at the steady state (Fig. 2d). The only region with relatively strong annual flows in the abyss is that in the southern part of the domain, that is, within or close to the latitudes of the circumpolar gap (Fig. 2d).

Fig. 2.

Annual horizontal circulation and deviations from the annual for January and July simulated by the idealized-basin model: (a)–(c) barotropic streamfunction and (d)–(f) velocity vectors at about 3.5 km. In (a)–(c) the contour interval is 3 Sv, with negative values shaded; in (d)–(f) the vectors represents mean velocities averaged for every second grid box in both directions.

Fig. 2.

Annual horizontal circulation and deviations from the annual for January and July simulated by the idealized-basin model: (a)–(c) barotropic streamfunction and (d)–(f) velocity vectors at about 3.5 km. In (a)–(c) the contour interval is 3 Sv, with negative values shaded; in (d)–(f) the vectors represents mean velocities averaged for every second grid box in both directions.

Superimposed on the annual circulation there is a strong seasonal cycle (Figs. 2b,c and 2e,f). Unlike the annual depth-integrated flows, the seasonal depth-integrated flows are mainly composed of the currents below the thermocline because baroclinic adjustment is not fast enough on seasonal time scales (Gill and Niiler 1973). The associated velocity values in the deep-oceanic interior are relatively large, even when simulated by a coarse-resolution model like the one employed for this illustration (Figs. 2e,f). These seasonal flows tend to be of opposite direction for opposite seasons. We shall illustrate in the next section that the deep-reaching seasonal currents simulated by a high-resolution model with realistic geometry have the same tendency.

Additional sensitivity experiments are conducted to illustrate that much of the seasonal variability in the abyss circulation is due to the seasonal variability in the wind field. When the annual cycle in the wind stress is suppressed, the deep seasonal flows vanish (not shown). In this case, the circulation corresponding to each individual month at the steady state very closely resembles the annual-mean circulation shown in Figs. 2a,d. If, instead, the annual cycle of wind stress is retained but the zonal component of annual-mean wind stress is removed from the forcing (i.e., τx = 0), the annual circulation essentially vanishes (Figs. 3a,d), whereas the deep-reaching seasonal flows persist (Figs. 3b,c and 3e,f). (Note that the arising curl resulting from the meridional component of annual wind stress ∂τy/∂x is retained in this sensitivity test; while relatively small, it may result in a nonzero depth-integrated annual circulation such as that seen in Fig. 3a. In addition, a weak thermohaline circulation may also develop because of the imposed small but finite value for vertical diffusivity.) It is also interesting to note that with the annual-mean zonal wind stress set to zero (and with vertical diffusivity being very low below the uppermost ocean), the strong circumpolar flow in the south essentially vanishes (Figs. 3a,d), as might be expected.

Fig. 3.

Same as in Fig. 2, but with annual-mean zonal wind stress removed from the model forcing.

Fig. 3.

Same as in Fig. 2, but with annual-mean zonal wind stress removed from the model forcing.

Outside of the tropics, the penetration of seasonal flows to the abyssal ocean has been related to the idea that the so-called vertical trapping scale (ze) of the ocean’s response to disturbances (Willebrand et al. 1980; Jayne and Marotzke 2001), which satisfies

 
formula

is a function of forcing frequency (ω) and wavenumber [k = (kx, ky)] (here, f is the Coriolis parameter, β is the meridional derivative of f, and N is the buoyancy frequency). This dependence of the trapping scale on forcing frequency and wavenumber for a typical midlatitude stratification is shown in Fig. 4 (cf. Willebrand et al. 1980). For a basinwide seasonal forcing this scale becomes comparable to, and even larger than, the mean depth of the ocean. The associated penetration of seasonal currents to the abyss implies an interaction with bottom topography. Such an interaction may turn out to be important for maintaining the enhanced rates of diapycnal mixing in the deep ocean in a manner, for example, similar to that proposed by Naveira Garabato et al. (2004) for the Southern Ocean. In the next section, we use results from an ocean model with resolution of ¼° to show that the wind-driven seasonal currents in the abyss may have velocities reaching locally 10−1 m s−1.

Fig. 4.

The vertical trapping scale (ze) given by Eq. (1) as a function of positive zonal wavenumber and frequency for β = 2 × 10−11 m−1 s−1, f = 1.2 × 10−4 s−1, and ky = 0.005 km−1. The buoyancy frequency is given by N(z) = Noexp (z/b), with No = 7.3 × 10−3 s−1 and b = 1.3 km. Shown are the lines of ze corresponding to 1, 2, and 5 km.

Fig. 4.

The vertical trapping scale (ze) given by Eq. (1) as a function of positive zonal wavenumber and frequency for β = 2 × 10−11 m−1 s−1, f = 1.2 × 10−4 s−1, and ky = 0.005 km−1. The buoyancy frequency is given by N(z) = Noexp (z/b), with No = 7.3 × 10−3 s−1 and b = 1.3 km. Shown are the lines of ze corresponding to 1, 2, and 5 km.

Another implication, which can be readily illustrated using models and evidence, which in fact has been presented using observations (Niiler and Koblinsky 1985), is that in the presence of topography the so-called bottom pressure torque can be an important term in the depth-integrated vorticity balance on seasonal time scales. Rough estimates can be used to support this suggestion. For example, assuming that even for the largest basinwide scales a representative horizontal velocity value for the near-bottom seasonal currents |ub| is 10−2 m s−1 (see next section), and using |D| = 10−3 as a representative value for the magnitude of large-scale bottom slopes, the associated near-bottom vertical velocity |wb| = |ub||D| can be as large as 10−5 m s−1 (such an estimate gives an upper limit for |wb|, because the cross-slope component of ub is likely to be smaller than |ub|). For comparison, a representative value for midlatitude wind stress curl is 10−7 N m−3 (e.g., Milliff et al. 2004), which gives a typical value for the Ekman pumping velocity of 10−6 m s−1.

The idealized numerical simulations of Anderson and Killworth (1977) show that the initial barotropic adjustment can be significantly influenced by topography. It is therefore expected that on seasonal time scales, bottom pressure torque can be an important term in the depth-integrated vorticity balance. To further support this suggestion, we use long-term control simulations from two global climate models. One of the models is a developmental version of the Canadian Centre for Climate Modeling and Analysis (CCCma) Coupled Model, version 3.5 (CanCM3.5), the initial version of which is described in Flato et al. (2000). This model is a further development of the CCCma climate model that is briefly described in section 4. The other model is one of the recent versions (version 2.1) of the GFDL Coupled Model (CM2.1; Delworth et al. 2006). In both models, the oceanic components have resolution of O(1°). The solutions analyzed here are from the corresponding control climate simulations obtained by running these global climate models for several centuries. Annual-mean and monthly mean fields are generated by averaging over two decades. Figure 5 shows the regions where the simulated near-bottom vertical velocity (i.e., the vertical velocity at the top of the bottom grid boxes) is relatively large, |wb| > 0.5 × l0−5 m s−1. For the annual-mean climates, the areas with such |wb| are mostly confined to the Southern Ocean and subpolar North Atlantic (Fig. 5, top panels); at lower latitudes, much of the annual-mean circulation is dominated by weaker |wb|, particularly in the GFDL model. The picture is markedly different for the simulated mean January circulation, wherein much of the low-latitude oceans are dominated by relatively large |wb| (Fig. 5, bottom panels).

Fig. 5.

Regions of relatively large near-bottom vertical velocity wb simulated by recent versions of the CCCma and GFDL coupled models corresponding to (top) mean annual climates and (bottom) mean January climates. Red areas correspond to wb > 0.5 × l0−5 m s−1, whereas green areas correspond to wb < −0.5 × 10−5 m s−1.

Fig. 5.

Regions of relatively large near-bottom vertical velocity wb simulated by recent versions of the CCCma and GFDL coupled models corresponding to (top) mean annual climates and (bottom) mean January climates. Red areas correspond to wb > 0.5 × l0−5 m s−1, whereas green areas correspond to wb < −0.5 × 10−5 m s−1.

Figure 6 presents the annual-mean and seasonal (January minus July) barotropic streamfunctions simulated by the CanCM3.5, and the corresponding Sverdrup streamfunctions for the subtropical gyres of the Pacific Ocean. It can be seen that, at least at this O(1°) resolution, much of the annual-mean depth-integrated flow in the interior of the gyres is in Sverdrup balance (Fig. 6, left panels). In contrast, the mean seasonal depth-integrated flows only marginally resemble the corresponding flows given by the Sverdrup solution (Fig. 6, right panels). One reason for this is, again, that on seasonal time scales, the vertical trapping scale ze becomes large enough for the interior wind-driven flows to penetrate to great depths. Essentially, the first baroclinic Rossby mode is not fast enough to confine seasonal flows to the thermocline. In the presence of favorable topographic features, much of the integrated seasonal circulation can be compensated away from the western boundary. In the subtropical North and South Pacific, for example, such topographic features are present near 145°E and near 175°W, respectively. The presence of these topographic features is reflected in the structure of the model seasonal streamfunction (Fig. 6, upper-right panel), but not in the corresponding Sverdrup streamfunction (Fig. 6, lower-right panel).

Fig. 6.

Mean (left) annual and (right) seasonal (January–July) (top) total and (bottom) Sverdrup transport streamfunctions. Red lines correspond to clockwise circulation, whereas blue lines correspond to counterclockwise circulation; contour interval is 10 Sv. The results are from the CanCM3.5.

Fig. 6.

Mean (left) annual and (right) seasonal (January–July) (top) total and (bottom) Sverdrup transport streamfunctions. Red lines correspond to clockwise circulation, whereas blue lines correspond to counterclockwise circulation; contour interval is 10 Sv. The results are from the CanCM3.5.

3. Global high-resolution model

In this section, strong seasonal variations in the deep-ocean circulation are illustrated using the Parallel Ocean Climate Model (POCM; Semtner and Chervin 1988, 1992; Stammer et al. 1996; McClean et al. 1997). The POCM is a primitive-equation model configured for the ocean between 75°S and 65°N with an average grid spacing of ¼°. It has 20 geopotential depth levels, ranging in thickness from 25 m in the upper ocean to 400 m in the abyssal ocean. The forcing consists of 3-day averages of wind stress fields and monthly surface heat fluxes, both derived from the European Centre for Medium-Range Weather Forecasts. The surface layer temperatures and salinities are restored to the Levitus and Boyer (1994) and Levitus et al. (1994) climatologies. The annual and monthly mean values for the ocean velocities used here were obtained by averaging from 1992 to 1998 of the POCM 4C integration.

Before we discuss the simulated seasonal flows in the abyss, there are some interesting features in the annual-mean circulation to be mentioned. Figure 7 presents the simulated by the POCM annual-mean currents in the deep northwest Pacific. The pattern shows some major deviations from what might be expected based on the Stommel and Arons (1960) model. These include the following: (a) the interior flows are more zonal than meridional, tending to group into jet-like structures; (b) at the subtropical latitudes, south of about 30°N, the boundary currents are fed by the interior fluxes, so that the latter have a predominant meridional direction toward the equator rather than toward the pole (this situation is summarized in Fig. 8a); and (c) the simulated meridional flows in the deep-oceanic interior can be significantly nonuniform, much like those observed. The latter point is illustrated in Fig. 8b using the model velocities together with the direct measurements of the deep circulation in the Brazil basin (Hogg and Owens 1999).

Fig. 7.

Mean annual circulation in the northwest Pacific at about 3.6-km depth simulated by the POCM. The scale arrow in the upper-left corner corresponds to 5 cm s−1. Light-gray regions correspond to ocean depths less than 5 km.

Fig. 7.

Mean annual circulation in the northwest Pacific at about 3.6-km depth simulated by the POCM. The scale arrow in the upper-left corner corresponds to 5 cm s−1. Light-gray regions correspond to ocean depths less than 5 km.

Fig. 8.

Zonal profiles of mean annual meridional velocities simulated by the POCM averaged (a) between 15°–25°N and below about 2.5 km in the North Pacific, and (b) between 15°–25°S and at about 2.5 km in the South Atlantic. Also shown in (b) are the mean meridional flows measured in the interior of the Brazil basin at the levels corresponding to the North Atlantic Deep Water (Hogg and Owens 1999).

Fig. 8.

Zonal profiles of mean annual meridional velocities simulated by the POCM averaged (a) between 15°–25°N and below about 2.5 km in the North Pacific, and (b) between 15°–25°S and at about 2.5 km in the South Atlantic. Also shown in (b) are the mean meridional flows measured in the interior of the Brazil basin at the levels corresponding to the North Atlantic Deep Water (Hogg and Owens 1999).

Much like in the case of the idealized basin model, there are significant seasonal variations in the abyssal ocean circulation simulated by the POCM, which are superimposed on the mean annual currents. In many places, these variations are of the order of 10−2 m s−1, with local values exceeding 5 × 10−2 m s−1. For a subregion in the northwest Pacific, the mean January and July current anomalies at about 3.6 km are presented in Fig. 9. These have, in many places, a direction opposite to each other (cf. Figs. 2e,f). Ideally, one would like to run sensitivity tests to identify the nature of this variability. However, based on our experiments with the idealized basin model in the previous section, it can be argued that much of the seasonal variability in the deep-ocean circulation simulated by the POCM can be attributed to the seasonal variations in the wind stress.

Fig. 9.

Mean monthly velocity anomalies simulated by the POCM in the subregion of the North Pacific at about 3.6-km depth: (top) January minus annual and (bottom) July minus annual. The scale arrow in the upper-left corner corresponds to 3 cm s−1. Light-gray regions correspond to ocean depths less than 5 km.

Fig. 9.

Mean monthly velocity anomalies simulated by the POCM in the subregion of the North Pacific at about 3.6-km depth: (top) January minus annual and (bottom) July minus annual. The scale arrow in the upper-left corner corresponds to 3 cm s−1. Light-gray regions correspond to ocean depths less than 5 km.

Jayne and Marotzke (2001) use a simple relation for the meridional velocity anomaly below the Ekman layer, such as

 
formula

to estimate seasonal overturning circulation in the ocean of depth H (x, y) (here, Δ denotes deviations from the corresponding annual-mean values). While there is no reason for this relation to hold locally, Jayne and Marotzke illustrate that outside of the equator it can describe the zonally integrated seasonal flows reasonably well. Thus, a change of sign in the zonally averaged Δτx (such as, e.g., in Fig. 1) would imply a change of sign in the corresponding Δυ, including in the deep ocean. In the interior, there is the corresponding reversal of zonal seasonal currents, with the magnitude of the deep flows strongly influenced by topography (Fig. 9). In general, however, the dynamics is governed by vorticity conservation, so relatively strong seasonal currents, including those in the abyss, can be generated locally without any changes in the local Ekman transport. For example, in our idealized basin model configuration, local wind stress curl has a strong seasonal cycle near 22°N and 22°S, whereas seasonal variations in the wind stress are small (Fig. 1). Hence, the net seasonal fluxes that compensate for the net seasonal Ekman fluxes across these latitudes are also small. However locally around 22°N and 22°S, the deep-reaching seasonal meridional and zonal flows can be as strong as elsewhere in the domain (Figs. 2e,f).

The largest seasonal variations in the deep-ocean circulation simulated by the POCM are found in the Indian Ocean. In this region, the currents during winter and summer have, in many places, a typical magnitude that is much larger than the magnitude of the annual currents (Fig. 10). These strong seasonal variations appear to be due to a combination of several factors. The Indian Ocean is characterized by a strong monsoonal cycle of winds (e.g., Trenberth et al. 1990; Peixoto and Oort 1992; Jayne and Marotzke 2001; see also next section). In addition, because the Ekman flows scale as sin−1(latitude), the same wind stress anomalies could result in stronger deep-reaching seasonal flows in the tropics compared to the extratropics.

Fig. 10.

Magnitude of horizontal velocity (cm s−1) at about 3 km simulated by the POCM in the Indian Ocean: (top) annual, (middle) January, and (bottom) July.

Fig. 10.

Magnitude of horizontal velocity (cm s−1) at about 3 km simulated by the POCM in the Indian Ocean: (top) annual, (middle) January, and (bottom) July.

A global view (north of 30°S) of the magnitude of mean currents at about 3-km depth is presented in Fig. 11. The most striking feature, again, is the presence of strong seasonal flows over vast areas of deep-oceanic interior. In addition to the Indian Ocean, relatively strong deep currents are simulated in the North Pacific, particularly in boreal winter. In some places, the magnitude of these seasonal flows is comparable to those along the western boundaries.

Fig. 11.

Magnitude of horizontal velocity (cm s−1) at about 3 km simulated by the POCM (between 30°S and 60°N): (top) annual, (middle) January, and (bottom) July.

Fig. 11.

Magnitude of horizontal velocity (cm s−1) at about 3 km simulated by the POCM (between 30°S and 60°N): (top) annual, (middle) January, and (bottom) July.

4. Projected changes

In this section we use a coupled climate model to evaluate potential changes in the mean annual cycle of wind stress over the ocean in response to plausible changes in the climate, and the associated changes in the seasonal circulation penetrating to the deep ocean. The model is one of the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) versions of the CCCma Coupled General Circulation Model, version 3 (CGCM3), the initial version of which is described in Flato et al. (2000). Briefly, the atmospheric component is a further development (Scinocca et al. 2008) of the spectral model described in McFarlane et al. (1992), configured here to have T63 resolution (≈2.8°). The oceanic component is derived from the GFDL MOM, version 1. There are two oceanic grid boxes per one atmospheric grid box in the zonal direction and three oceanic grid boxes per one atmospheric grid box in the meridional direction. Mesoscale oceanic eddies are parameterized using the Gent and Mc Williams (1990) mixing scheme. Sea ice thermodynamics is governed by an energy balance model (McFarlane et al. 1992), whereas sea ice dynamics uses a cavitating-fluid rheology (Flato and Hibler 1992). Here, we analyze two climate simulations by the model. One of them is the control simulation, wherein the external forcing parameters, such as the concentration of atmospheric CO2, are kept fixed at their preindustrial level. In the other simulation, the model is forced using the historical record for concentration of greenhouse gases (GHGs) and, for the twenty-first century, using the IPCC Special Report on Emissions Scenarios (SRES) A1B scenario to specify the concentration of GHGs. The discussed quantities represent the mean climate conditions simulated by the end of the twenty-first century and averaged over a 10-yr period between 2091 and 2100.

The seasonal cycle of wind stress is sometimes described by the difference between the mean January and mean July conditions (Jayne and Marotzke 2001). This field, both corresponding to the preindustrial climate conditions and to the climate projected by the end of the twenty-first century, is shown in Fig. 12. For individual ocean basins, the projected changes in the seasonal wind stresses are quite large (Fig. 12). This is the case, for example, in the extratropical North Pacific during boreal winter. It is therefore expected, based on the discussion in the previous sections, that if these projected changes in the annual cycle of wind stress were to be realized, the seasonal circulation at the deepest parts of the ocean could change accordingly. To predict the latter, particularly if one were interested in details, one possibility is to run a high-resolution ocean model, such as the POCM. This, however, is beyond the scope here. Instead, in what follows, we consider the changes in an integral ocean quantity, such as seasonal overturning circulation. The hope here is that, while locally the climate model velocities might be inaccurate because of the model’s resolution, some of the integral quantities might be simulated better. Furthermore, we narrow this discussion focusing only on the North Pacific. This is mainly for two reasons: First, the projected changes in the seasonal wind stress are relatively large there, which, combined with the width of the basin, results in large changes in the zonally integrated Ekman transports. Second, it is hoped that much of the North Pacific overturning and its response to changes in the winds can be captured without the need to employ an eddy-resolving model.

Fig. 12.

Seasonal cycle (January minus July) of wind stress over the ocean simulated by the CCCma CGCM3 T63 and corresponding to (top) the preindustrial climate conditions and (bottom) the climate conditions projected by the end of the twenty-first century using the IPCC SRES A1B scenario to specify the concentration of greenhouse gases. The vectors are plotted for every second atmospheric grid box in the model. The scale arrow corresponds to 0.5 Pa.

Fig. 12.

Seasonal cycle (January minus July) of wind stress over the ocean simulated by the CCCma CGCM3 T63 and corresponding to (top) the preindustrial climate conditions and (bottom) the climate conditions projected by the end of the twenty-first century using the IPCC SRES A1B scenario to specify the concentration of greenhouse gases. The vectors are plotted for every second atmospheric grid box in the model. The scale arrow corresponds to 0.5 Pa.

There are no major sources of deep-water formation in the North Pacific. The simulated annual-mean overturning resulting from meridional Ekman transports is confined to the upper ocean, as expected (not shown). However, for individual months the overturning anomaly (relative to the annual mean) penetrates to the abyssal ocean (Fig. 13; cf. Bryan 1982; Boning and Herrmann 1994; Nakano et al. 1999; Jayne and Marotzke 2001). Moreover, it is projected to change, for some months significantly. Examples of such changes for January and October in the extratropical North Pacific are shown in Fig. 13 (lower panels). For both of these months, the cells associated with the deep-reaching overturning anomalies centered around 35° and 50°N are projected to intensify (Fig. 13). As expected, these changes in the monthly overturning anomalies can be related to the changes in the meridional Ekman transport anomalies across representative latitudes (Fig. 14). In particular, the southward (northward) Ekman transport anomaly across 37°N during January (October) is projected to roughly double (Fig. 14, left panel), consistent with the changes in the corresponding overturning anomaly derived from the 3D model ocean velocities (Fig. 13). Across 50°N, the northward (southward) Ekman transport anomaly also increases during January (October) (Fig. 14, right panel), which is consistent with the changes in the corresponding overturning anomaly centered near this latitude (Fig. 13). Thus, surface wind stress data can be effectively used to identify the seasons and ocean basins, with the largest changes in the deep-reaching overturning cells in response to changing climate conditions.

Fig. 13.

Mean monthly anomalies of the meridional overturning circulation in the extratropical North Pacific (relative to the corresponding annual-mean overturning) and representing (top) the preindustrial climate conditions and (bottom) the climate conditions projected by the end of the twenty-first century: (right) January and (left) October. Contour interval is 2 Sv. The shaded overturning cells correspond to counterclockwise circulation anomalies.

Fig. 13.

Mean monthly anomalies of the meridional overturning circulation in the extratropical North Pacific (relative to the corresponding annual-mean overturning) and representing (top) the preindustrial climate conditions and (bottom) the climate conditions projected by the end of the twenty-first century: (right) January and (left) October. Contour interval is 2 Sv. The shaded overturning cells correspond to counterclockwise circulation anomalies.

Fig. 14.

Monthly mean anomalies of meridional Ekman transports (relative to the corresponding annual-mean transports) across (left) 37°N and (right) 50°N in the Pacific and representing (circles) the simulated mean preindustrial climate conditions and (triangles) the mean climate conditions projected by the end of the twenty-first century (2091–2100).

Fig. 14.

Monthly mean anomalies of meridional Ekman transports (relative to the corresponding annual-mean transports) across (left) 37°N and (right) 50°N in the Pacific and representing (circles) the simulated mean preindustrial climate conditions and (triangles) the mean climate conditions projected by the end of the twenty-first century (2091–2100).

Overall, it is confirmed that the large-scale seasonal flows in the abyssal ocean are subject to change in response to changes in the seasonal cycle of surface wind stress. Based on the high-resolution model results in the previous section, it might be expected that locally the changes in the deep-reaching seasonal currents could be much larger than those implied by the changes in the corresponding integrated circulation.

5. Discussion and conclusions

Using a set of models, including one with a resolution of ¼°, we consider several aspects of the seasonal currents in the deep ocean. Previously, it was noted (Jayne and Marotzke 2001) that while the amplitude of the seasonal overturning circulation can be as large as 50 Sv, the associated velocities in the deep ocean are small. Jayne and Marotzke estimate the corresponding horizontal and vertical velocities to be, respectively, of the order of 10−3 m s−1 and 10−6 m s−1. However, these estimates apply only to zonally averaged circulation. Locally, seasonal currents can be much stronger, directed often in opposite directions for opposite seasons. Furthermore, we show that over vast areas of the deep interior, particularly in the Indian Ocean, annual-mean circulation represents a small residual of much stronger seasonal flows. In many places, the seasonal horizontal velocities are of the order of 10−2 m s−1, reaching 10−1 m s−1 locally; the corresponding vertical velocities can reach values of the order of 10−5 m s−1, even in coarse-resolution climate models. We employ an idealized geometry model to show that much of this seasonal variability in the deep-ocean circulation can be attributed to annual cycle of wind stress, combined with the significant increase in the vertical trapping depth for basin-scale seasonal forcing.

There can be several interesting implications of this phenomenon. First, with the near-bottom vertical velocity reaching values of the order of 10−5 m s−1 in many places, it is reasonable to expect that the so-called bottom pressure torque can be an important term in the depth-integrated vorticity balance on seasonal time scales. To support this suggestion we use climate models and show that in the presence of topography, the integrated seasonal flows in the interior are not in Sverdrup balance, at least not in its original form (Sverdrup 1947), as might be expected from the earlier theories (e.g., Gill and Niiler 1973; Willebrand et al. 1980). It may be, however, that the so-called generalized Sverdrup balance, which is given by (e.g., Willebrand et al. 1980; in conventional notation)

 
formula

is a better approximation for vertically integrated transport in the ocean [U = (U, V)] on seasonal time scales. This vorticity balance implies that the net mass flux across the contours of constant f /H is balanced by the vorticity input resulting from wind stress curl. After rewriting it as

 
formula

it can be more readily seen how the topographic effects modify the “original” Sverdrup balance. Limited observations seem to support this generalized balance (Niiler and Koblinsky 1985).

Second, an interaction of strong flows with bottom topography may contribute to diapycnal mixing in the deep ocean. For example, under certain conditions a flow over topography can generate internal waves, the breaking of which may contribute to sustaining the enhanced level of turbulence in the deep ocean. An idealized case of internal waves generated this way is discussed in Gill (1982, 142–146, 268–274). For a deep flow with velocity U, the character of the waves depends on the ratio kU/N, with k−1 being the characteristic scale of the bottom topography. For kU/N > 1, the waves decay exponentially with height, so that the energy is trapped near the bottom; for kU/N < 1, the waves can propagate, transferring their energy to remote levels.

As reviewed by Wunsch and Ferrari (2004), it is the flows of tidal nature, rather than wind-driven currents, that are thought to be the most efficiently placed for deep mixing. This is because wind-driven motions are generally thought to be weak in the deep-oceanic interior (except maybe in the Southern Ocean). We show that this may only apply to annual-mean circulation, whereas on seasonal time scales the relatively strong wind-driven flows can penetrate to the ocean floor over vast areas, particularly in the Indian Ocean. Thus, while speculative, there is a possibility that these deep-reaching seasonal flows over topography could be an important source of internal wave energy in the abyss. In the Southern Ocean, for example, where relatively strong large-scale and mesoscale geostrophic flows reach to the bottom, the observed enhanced level of turbulence in the areas of the so-called hot spots has been interpreted as being primarily sustained by internal wave generation as these nontidal flows impinge upon bottom topography (Naveira Garabato et al. 2004).

Another related topic touched here concerns projected changes in the annual cycle of wind stress. Using a climate model, it is found that under a plausible climate change scenario, the amplitude of the mean annual cycle of wind stress may change. Among the regions where such changes are most pronounced, such as the extratropical North Pacific, it is shown that the data on surface wind stress can be effectively used to identify the seasons with the largest changes in the deep-reaching overturning cells. Locally, the corresponding changes in the seasonal currents in the abyss can be much larger than those implied by the changes in the zonally integrated circulation.

Acknowledgments

I thank Robin Tokmakian for providing the POCM data used in this study and Alex Sen Gupta for comments and his help with the POCM data. I also thank the Geophysical Fluid Dynamics Laboratory for making their model output files publicly accessible. Comments from Bill Merryfield and reviewers helped to improve the manuscript.

REFERENCES

REFERENCES
Anderson
,
D. L. T.
, and
A. E.
Gill
,
1975
:
Spin-up of a stratified ocean with application to upwelling.
Deep-Sea Res.
,
22
,
583
596
.
Anderson
,
D. L. T.
, and
P. D.
Killworth
,
1977
:
Spin-up of a stratified ocean, with topography.
Deep-Sea Res.
,
24
,
709
732
.
Boning
,
C. W.
, and
P.
Herrmann
,
1994
:
Annual cycle of poleward heat transport in the ocean: Results from high-resolution modeling of the North and equatorial Atlantic.
J. Phys. Oceanogr.
,
24
,
91
107
.
Brink
,
K. H.
,
1989
:
Evidence for wind-driven current fluctuations in the western North Atlantic.
J. Geophys. Res.
,
94
,
2029
2044
.
Bryan
,
K.
,
1982
:
Seasonal variation in meridional overturning and poleward heat transport in the Atlantic and Pacific Oceans: A model study.
J. Mar. Res.
,
40
, (
Suppl.
).
39
53
.
Delworth
,
T. L.
, and
Coauthors
,
2006
:
GFDL’s CM2 global coupled climate models. Part I: Formulation and simulation characteristics.
J. Climate
,
19
,
643
674
.
Flato
,
G. M.
, and
W. D.
Hibler
III
,
1992
:
Modeling pack ice as a cavitating fluid.
J. Phys. Oceanogr.
,
22
,
626
651
.
Flato
,
G. M.
,
G. J.
Boer
,
W. G.
Lee
,
N. A.
McFarlane
,
D.
Ramsden
,
M. C.
Reader
, and
A. J.
Weaver
,
2000
:
The Canadian Centre for Climate Modelling and Analysis global coupled model and its climate.
Climate Dyn.
,
16
,
451
467
.
Gent
,
P. R.
, and
J. C.
McWilliams
,
1990
:
Isopycnal mixing in ocean general circulation models.
J. Phys. Oceanogr.
,
20
,
150
155
.
Gill
,
A. E.
,
1982
:
Atmosphere-Ocean Dynamics.
Academic Press, 662 pp
.
Gill
,
A. E.
, and
P.
Niiler
,
1973
:
The theory of seasonal variability in the ocean.
Deep-Sea Res.
,
20
,
141
177
.
Hogg
,
N. G.
, and
W. B.
Owens
,
1999
:
Direct measurement of the deep circulation within the Brazil Basin.
Deep-Sea Res. II
,
46
,
335
353
.
Jayne
,
S. R.
, and
J.
Marotzke
,
2001
:
The dynamics of ocean heat transport variability.
Rev. Geophys.
,
39
,
385
411
.
Koblinsky
,
C. J.
, and
P. P.
Niiler
,
1982
:
The relationship between deep ocean currents and winds east of Barbados.
J. Phys. Oceanogr.
,
12
,
144
153
.
Levitus
,
S.
, and
T.
Boyer
,
1994
:
Temperature.
Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 117 pp
.
Levitus
,
S.
,
R.
Burgett
, and
T.
Boyer
,
1994
:
Salinity.
Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 99 pp
.
McClean
,
J. L.
,
A. J.
Semtner
, and
V.
Zlotnicki
,
1997
:
Comparisons of mesoscale variability in the Semtner-Chervin 1/4 degree model, the Los Alamos Parallel Ocean Program 1/6 degree model, and TOPEX/POSEIDON data.
J. Geophys. Res.
,
102
,
25203
25226
.
McFarlane
,
N. A.
,
G. J.
Boer
,
J-P.
Blanchet
, and
M.
Lazare
,
1992
:
The Canadian Climate Centre second-generation general circulation model and its equilibrium climate.
J. Climate
,
5
,
1013
1044
.
Milliff
,
R. F.
,
J.
Morzel
,
D. B.
Chelton
, and
M. H.
Freilich
,
2004
:
Wind stress curl and wind stress divergence biases from rain effects on QSCAT surface wind retrievals.
J. Atmos. Oceanic Technol.
,
21
,
1216
1231
.
Nakano
,
H.
,
R.
Furue
, and
N.
Suginohara
,
1999
:
Effect of seasonal forcing on global circulation in a world ocean general circulation model.
Climate Dyn.
,
15
,
491
502
.
Naveira Garabato
,
A. C.
,
K. L.
Polzin
,
B. A.
King
,
K. J.
Heywood
, and
M.
Visbeck
,
2004
:
Widespread intense turbulent mixing in the Southern Ocean.
Science
,
303
,
210
213
.
Niiler
,
P. P.
, and
C. J.
Koblinsky
,
1985
:
A local time-dependent Sverdrup balance in the eastern North Pacific Ocean.
Science
,
229
,
754
756
.
Niiler
,
P. P.
,
J.
Filloux
,
W. T.
Liu
,
R. M.
Samelson
,
J. D.
Paduan
, and
C. A.
Paulson
,
1993
:
Wind-forced variability of the deep eastern North Pacific: Observations of seafloor pressure and abyssal currents.
J. Geophys. Res.
,
98
,
22589
22602
.
Peixoto
,
J. P.
, and
A. H.
Oort
,
1992
:
Physics of Climate.
American Institute of Physics, 520 pp
.
Qiu
,
B.
,
2002
:
Large-scale variability in the midlatitude subtropical and subpolar North Pacific Ocean: Observations and causes.
J. Phys. Oceanogr.
,
32
,
353
375
.
Scinocca
,
J. F.
,
N. A.
McFarlane
,
M.
Lazare
,
J.
Li
, and
D.
Plummer
,
2008
:
The CCCma third generation AGCM and its extension into the middle atmosphere.
Atmos. Chem. Phys.
,
8
,
7883
7930
.
Semtner
,
A. J.
, and
R. M.
Chervin
,
1988
:
A simulation of the global ocean circulation with resolved eddies.
J. Geophys. Res.
,
93
,
15502
15522
.
Semtner
,
A. J.
, and
R. M.
Chervin
,
1992
:
Ocean general circulation from a global eddy-resolving model.
J. Geophys. Res.
,
97
,
5493
5550
.
Stammer
,
D.
,
R.
Tokmakian
,
A. J.
Semtner
, and
C.
Wunsch
,
1996
:
How well does a 1/4 degree global circulation model simulate large-scale oceanic observations?
J. Geophys. Res.
,
101
,
25779
25811
.
Stommel
,
H. M.
, and
A. B.
Arons
,
1960
:
On the abyssal circulation of the world ocean—I. Stationary planetary flow patterns on a sphere.
Deep-Sea Res.
,
6
,
140
154
.
Sverdrup
,
H.
,
1947
:
Wind-driven currents in a baroclinic ocean; with application to the equatorial currents of the eastern Pacific.
Proc. Natl. Acad. Sci. USA
,
33
,
318
326
.
Trenberth
,
K. E.
,
W. G.
Large
, and
J. G.
Olson
,
1990
:
The mean annual cycle in global ocean wind stress.
J. Phys. Oceanogr.
,
20
,
1742
1760
.
Weaver
,
A. J.
, and
Coauthors
,
2001
:
The UVic Earth System Climate Model: Model description, climatology and application to past, present and future climates.
Atmos.–Ocean
,
39
,
361
428
.
White
,
W. B.
,
1978
:
A wind-driven model experiment of the seasonal cycle of the main thermocline in the interior midlatitude North Pacific.
J. Phys. Oceanogr.
,
8
,
818
824
.
Willebrand
,
J.
,
S. G. H.
Philander
, and
R. C.
Pacanowski
,
1980
:
The oceanic response to large-scale atmospheric disturbances.
J. Phys. Oceanogr.
,
10
,
411
429
.
Wunsch
,
C.
, and
R.
Ferrari
,
2004
:
Vertical mixing, energy, and the general circulation of the oceans.
Annu. Rev. Fluid Mech.
,
36
,
281
314
.

Footnotes

Corresponding author address: Oleg A. Saenko, CCCma, 3964 Gordon Head Road, Victoria, BC V8N 3X3, Canada. Email: oleg.saenko@ec.gc.ca