Semyon A. Grodsky and James A. Carton

July 2, 2002

Revised August 289, 2002

senya@atmos.umd.edu

carton@atmos.umd.edu

pathways.doc and pictures fig#.eps

Department of Meteorology, University of Maryland, College Park, MD 20742

Abstract

Subtropical cells connect subduction zones of the eastern subtropics of both hemispheres to the equatorial current systems via equatorward flow in the thermocline. Some of this thermocline water is converted back into warm near-surface water in the eastern equatorial cold tongue from whenceich it is then exported through zonal and meridional divergence. Here we examine the export pathways from the cold tongue in the Atlantic based on recently available near-surface drifter observations. We find that, similar to its Pacific counterpart, water upwelled in the Atlantic cold tongue follows multiple pathways back into the subtropics, but not directly back to the subduction zones. Thus the subtropical cells are open to extensive influence from, and exchange with, the subtropical and midlatitude ocean.

The temperature and circulation of the tropical Atlantic are both strongly seasonal. In boreal spring warm water spans the equator to 15^oN. Within a few degrees of the equator extending to 10^oS lie the northern and southern branches of the westward flowing South Equatorial Current. The northern branch of the South Equatorial Current provides source water for the near-coastal North Brazil Current system and its associated rings and eddies (Richardson and Walsh, 1986; Fratantoni et al., 2000). By summer in response to a northward shift of the northeast and southeast trade wind systems the eastward North Equatorial Countercurrent forms. Along the equator mixed layer temperatures drop by several degrees in the region 30^oW-0^oW, 4^oS-2^oN in response to a shallowing thermocline and strengthening easterlies, forming the center of the equatorial cold tongue.

In zonal and time average the meridional-vertical circulation of the upper ocean is dominated by northern and southern subtropical cells (STCs) in which high salinity water is subducted in the subtropics, flows equatorward along the thermocline, then upwells into the mixed layer of the equatorial zone, where it is exported poleward (Bryan, 1991; Gu and Philander, 1997). If this water is then returned to the subduction regions it forms a closed loop with a cycle time of a few years. However, examination of the pathways of water in the upper limb of the STCs in the Pacific by Johnson (2001) shows northwestward transport supported mostly by the Ekman drift. Hence this water reenters the subtropics in the west, away from the subduction zones that feed the lower limb of the STCs. This result suggests the STCs are open circulations, rapidly exchanging water with the ocean's subtropical gyres. Motivated by this result, here we examine the pathways of mixed layer water as observed in the near-surface drifter record of the tropical Atlantic during 1997-2001 and in related data sets.

Our analysis uses 55 drifter trajectories within the 15^oS - 30^oN band, obtained from the WOCE/TOGA archive at the Atlantic Ocean Marine Laboratory (NOAA/AOML), that pass within the rectangular domain 4^oS-2^oN and 30^oW-0^oE during the cold tongue seasons of boreal summer and fall (June-November when SSTs are low and entrainment is strong). No attempt was made to resolve differences within the cold tongue seasons because of data limitations. Many of the trajectories were obtained during 1999-2000 peaking at ~1000 buoy day yr^-1 of observation and decreasing to about half of that level in 2001. No drifters were available from years prior to 1997. A few drifters which crossed the cold tongue were available from the Programme Francais Ocean et Climat Dans l'Atlantique Equatorial and Seasonal Response of the Atlantic Ocean Experiment (FOCAL/SEQUAL, Reverdin and McPhaden, 1986) of 1983-1984. We chose to exclude them from this analysis because of the difficulty in accounting for changes in drogue design and because of the unusual oceanographic conditions of 1983-4 due to El Nino. Experiments with an ocean model reanalysis (Carton et al., 2000) show that the period 1997-20001is sufficient to estimate has a stable seasonal cycle in the tropical Atlantic. During this period the most unusual aspect is that equatorial currents were unusually stronger during the first half of 1997. However, only 10% of drifter data used was collected during 1997 (see inlay to Fig. 2a).

Because of limitations on the number and coverage of surface drifters, we supplement our velocity discussion using a multivariate optimal interpolation analysis of climatological seasonal near-surface currents on a 3^ox2^o grid following Grodsky and Carton (2001a,b) and Grodsky et al. (2001) (see Fig. 1). The analysis combines Eulerian velocity estimates obtained from the drifter tracks, following the methodology of Hansen and Poulain (1996), with near-surface pressure variations based on TOPEX/POSEIDON altimetry (Koblinsky, personal communication, 1997) and ERS 1/2 monthly scatterometer winds of Bentamy et al. (2001), which together provide information on the geostrophic and ageostrophic components of the near-surface currents. A first guess estimate of the seasonal currents is provided by historical ship drifts (available on the Ocean Current Drifter Data CD-ROM provided by NOAA/NODC). The advantages and disadvantages of the ship drift data have been discussed by Richardson and Walsh (1986). Further details are provided in Grodsky and Carton (2001a,b) and Grodsky et al. (2001). The ocean mixed layer depth used in this study is the seasonal climatology of Kara et al. (2000).
 
 

Fig. 1 Climatological nearsurface currents. (a) Annual average nearsurface currents and entrainment/detrainment velocity, we. (b) Zonally averaged (35oW - 10oE) we with latitude. Note the presence of detrainment near 5oN. (c) Seasonal cycle of we with latitude. (d) Seasonal cycle of Ekman pumping with latitude. Ekman pumping close to the equator is calculated using the linear friction balance All contours are drawn at the intervals [-7.5 -2.5 2.5 7.5 12.5 17.5]x10-6 ms-1. Positive (negative) values are shown with in solid (dashed) lines.

We estimate the rate of entrainment of water into the mixed layer, w_e, following Swenson and Hansen (1999) as the difference between sum of the rate of deepening of the mixed layer, dh/dt, and upwelling at its base, which can be calculated from the divergence of mixed layer velocity, -hdivu. While entrainment heat flux is only defined while w_e is upward, actually w_e can be either upward or downward (entrainment or detrainment). Our estimates in Fig. 1b show that on annual average entrainment zonally averaged between 35^oW and 10^oE is mainly confined to a band of latitudes between 76^oS-26^oN, with a maximum of ~12x10^-6 m/s at 2^oS and aband average of ~(5.92 ± 10.54)x10^-65 ms^-1(1.7 ± 0.3 m dy^-1),thea values similar to those found by Weisberg and Qiao (2000) and by Meinen et al. (2001) in the Pacific cold tongue. Entrainment also occurs near the western boundary but our estimates become unreliable in this region because of lack of data. The maximum entrainment is displaced slightly south of the equator in the east as a result of the southerly component of wind that produces upwelling to the south and downwelling to the north of the equator (Perigaud et al., 1997). The uncertainties here and below derive from estimates of velocity accuracy (Grodsky and Carton, 2001a) and data distribution.

Integrating over the region of upward w_e within the meridionalzonal band 76^oS-26^oN in the ocean in the longitude band of the cold tongue 350^oW-10^oEgives a total upwelling estimate of (2631 ± 8)x10^-6 m³s^-1 (2631 Sv) about half of the value found by Johnson (2001) in the equatorial Pacific. Our upwelling estimate is in line with the Roemmich (1983) estimate of 25 Sv obtained based on the Ekman divergence between 8^oS and 8^oN. Weak off equatorial detrainment with peak velocity of -3.5x10^-6 ms^-1at ~5^oN occurs mostly between North of 35^oN and 8^oN (Fig. 1b)and could be evidence of the downwelling limb of the shallow tropical cell (Lu et al., 1998). In distinction from the Pacific (Johnson, 2001), the shallow tropical cell in the Atlantic develops only north of the equator. weak 3.5x10^-6 ms^-1(0.3 m dy^-1) detrainment occurs (Fig. 1b). Integration over the zonal band between 350^oWN and 10^oENwhere detrainment occurs gives a total downwelling rate of 49 ± 26 Sv suggesting that roughly 15%quarter of the water that upwells on the equator downwells before it reaches 810^oN. Note that Richardson and Reverdin (1987) Our estimate of this found a stronger downwelling ratecompares well to that of 8Sv of Richardson and Reverdin (1987) based on ship drift observations. Finally, we consider seasonal variations of w_e. Entrainment as well as detrainment increases in vicinity of the equator during boreal summer and fall (Fig.1c). These increases correspond to the seasons of peak Ekman pumping (Fig. 1d). This correspondence reflects the principal role of winds in driving w_e. 

Next we consider the fate of the upwelled water by examining 55 drifter trajectories that pass through the cold tongue region (30^oW-0^oE, 4^oS-2^oN) during the cold tongue seasons (Fig. 2a). The time of origin of the trajectories is distributed inhomogeneously between different months and years (see insets in Fig. 2a). Most frequently the trajectories originate during late summer-early autumn and during 1999-2000. We partition the drifter trajectories based on whether a drifter finally crosses the meridian 40^oW and if it does, whether it crosses this meridian north or south of latitude 5^oN.Drifters that remain east of 40^oW are selected separated into two groups by checking if a drifter eventuallygoes finally south of the equator. As we shall see, tThispartitioning allows us to discriminate between trajectories entering the North Equatorial Current, the near-coastal North Brazil Current system, remaining in the southern branch of the South Equatorial Current, or entering the eastward North Equatorial Countercurrent. The trajectory points of origin are defined as the location of deployment or the location where a drifter trajectory crosses into the cold tongue region for thea first time. The results are presented in Figs. 2b-e, along with a mean trajectory for each case. 
 
 

Fig. 2 Fate of drifters crossing the cold tongue (4oS-2oN, 30oW-0oE) during the six month period June-November. (a) Combined drifter trajectories. Inlays in (a) present distribution of points of origin by months and by years. (b) Drifter trajectories entering the North Equatorial Current. (c) Drifter trajectories entering the South Equatorial Current and the North Brazil Current system. (d) Drifter trajectories entering the South Equatorial Current and continuing in the Southern Hemisphere. (e) Drifters entering the North Equatorial Countercurrent and thus being carried eastward. Mean trajectory of each group is shown by solid line with open circles indicating four-month intervals. Shaded rectangles show standard deviation of drifter coordinates during a month. White rectangles show the initial spread of drifter locations belonging to each group. Digits Dr indicate number of drifters in each group while Mn indicates number of months when at least 2 drifters are alive.

 
 

A third group of 11 buoys shown in Fig. 2d that is transported into the southern hemisphereof 5^oN originates south of the equator and east of 20^oW. These trajectories are swept southwestward by the southern branch of the South Equatorial Current (the SEC group) and reach 10^oS after approximately twoone years.

The remaining 23 trajectories, almost 40% of the total, originate close to or north of the equator, mostly east of 20^oW. These trajectories remain in the eastern basin as a result of rapidly entering the eastward North Equatorial Countercurrent in the second half of the year (Fig. 2f). Tropical Instability Waves, which perturb the meridional position of the North Equatorial Countercurrent, produce substantial divergence in longitudinal position of this group (the NECC group). However, on average they require one year of travel to approach the Greenwich meridian implying an average eastward speed of less than 10 cm/s.

The results presented above suggest that the point of origin within the cold tongue has a significant influence on the near-surface trajectories because of the direction and seasonality of the prevailing currents. We explore this possibility by calculating simulated trajectories using the gridded analysis of seasonal near-surface currents described in Section 2 (see Fig. 3). All 28 simulated trajectories begin in August, which is a peak month of the cold tongue. 30% of the simulated trajectories originating in the northwestern quadrant of the cold tongue (30^oW-20^oW, 0^oN-2^oN) and even south of the equator between 2015^oW and 150^oW go northwestward (like the NEC group). 10% of the trajectories beginning on or south of the equator on the western periphery of the cold tongue are transported by the northern branch of the South Equatorial Current westward into the North Brazil Current, and tthence into the northern subtropical gyre (SEC+NBC group) northwestward. 20% of the water parcelssimulated trajectories with a point of origin along the southern edge of the cold tongue region between 25^oW and 0^oE, in contrast, follow the southern branch of the South Equatorial Current into the southern subtropics (SEC group). 40% of the simulated trajectories that originate Water parcels in the northwest part 40% of the cold tongue region rapidly enter the North Equatorial Countercurrent and thus are transported eastward into the Gulf of Guinea (like the NECC group). Calculations of trajectories beginning earlier or later during the upwelling season reveal the same four groups of trajectories but with different partitioning percentages.
 
 

Fig. 3 Average particle trajectories with points of origin in the cold tongue (4oS-2oN, and 30oW-0oE) during August accounting for the seasonally varying currents. Trajectories shown here have points of origin on a 5o and 2o longitude-latitude grid. Four symbols mark trajectories corresponding to Fig 2b,c,d,e. The trajectories are overlaid on climatological August SST obtained from Reynolds and Smith (1994).

 
 

Recent studies point to the eastern northern and particularly southern subtropics as major subduction regions for water entering the tropical Atlantic thermocline (Zhang et al., 2002) and providing a source for the 3126 Sv we find being entrained into the mixed layer in the cold tongue region of the eastern equatorial Atlantic during boreal summer and fall. However, despite its potentially important role in Atlantic climate the surface limb of the Atlantic subtropical cells have received limited attention (with the exception of Reverdin and McPhaden 1986). In this paper we examine the trajectories of 55 near-surface drifters deployed over the past five years and find four major pathways for water exiting the cold tongue region, none of which lead directly back to the subduction regions.

The drifters originating on the western periphery of the cold tongue (west of ~15^oW) are transported into the northern subtropical gyre by the westward moving northern branch of the South Equatorial Current either directly across the North Equatorial Countercurrent in boreal spring, or through the near-coastal North Brazil Current system. The second pathway is the faster route to the subtropics and requires roughly one and a half years, while the first requires approximately two years.

The drifters that have a point of origin on or south of the equator, east of 20^oW follow a third pathway southwestward in the southern branch of the South Equatorial Current reaching the southern subtropical gyre after about two years. The drifterswhose point of originlies east of 20^oW mostly north of the equator are diverted eastward by the North Equatorial Countercurrent and are carried into the Gulf of Guinea, reaching the Greenwich meridian after a year.

Given the small number of drifters along with their inhomogeneous spatial and temporal distribution, we can't quantify relative importance of the pathways they identify. The gridded velocity data are somewhat helpful, but their lack of subseasonal variability may be a problem. The difference between the numbers obtained from drifter trajectories and gridded velocities may reflect the limited number and spatial and temporal inhomogeneity of drifters, or it may reflect the lack of subseasonal variability in the analysis of currents. Some additional information may be obtained from analysis of realistic general circulation model simulations. Further narrowing of the uncertainty of the observational estimates will require additional near-surface drifter deployments in the tropical Atlantic cold tongue region, perhaps as part of a larger effort to define the climate variability of the tropical Atlantic. 

Acknowledgements

Bryan, K., Poleward heat transport in the ocean, Tellus, 43, 104-115, 1991.

Carton, J.A., G. Chepurin, X. Cao, and B. Giese, A simple Ocean Data Assimilation analysis of the global upper ocean 1959-1995, Part 1: Methodology, J. Phys. Oceanogr., 30, 294-309, 2000.

Grodsky, S.A. and J.A. Carton. Intense surface currents in the Tropical Pacific during 1996-1998, J. Geoph. Res., 106, 16,673-16,684, 2001a.

Grodsky, S.A. and J.A. Carton, and R. Murtugudde, Anomalous surface currents in the tropical Indian Ocean, Geoph. Res. Lett., 28, 4207-4210, 2001b.

Gu, D.F. and S.G.H. Philander, Interdecadal climate fluctuations that depend on exchanges between the tropics and extratropics, Science, 275, 805-807, 1997.

Hansen, D.V. and P.-M. Poulain, Quality control and interpolations of WOCE-TOGA drifter data, J. Atmos. Oceanic Technol., 13, 900-909, 1996.

Fratantoni, D.M., W.E. Johns, T.L. Townsend, and H.E. Hulburt, Low-level circulation and mass transport pathways in a model of the tropical Atlantic ocean, J. Phys. Oceanogr., 30, 1944-1966, 2000.

Johnson, G.C., and M.J. McPhaden, Interior picnocline flow from the subtropical to the equatorial Pacific ocean, J. Phys.Oceanogr., 29, 3073-3089, 1999.

Johnson, G.C., The Pacific ocean subtropical cell surface limb, Geoph. Res. Lett., 28, 1771-1774, 2001.

Kara, A.B., P.A. Rochford, and H.E. Hurlburt, An optimal definition for ocean mixed layer depth, J. Geoph. Res., 105, 16,803-16,621, 2000.

Kleeman, R., J.P. McCreary, and B.A. Klinger, A mechanism for generating ENSO decadal variability, Geoph. Res. Lett., 26, 1743-1746, 1999.

Lu, P., J.P. McCreary, Jr., and B.A. Klinger, Meridional circulation cells and the source waters of the Pacific Equatorial Undercurrent, J. Phys. Oceanogr., 28, 62-84, 1998.

McCreary, J.P. and P. Lu, Interaction between the subtropical and equatorial ocean circulations: the subtropical cell, J. Phys.Oceanogr., 24, 466-497, 1994.

Malanotte-Rizzoli P., K. Hedstrom, H. Arango, and D.B. Haidvogel, Water mass pathways between the subtropical and tropical ocean in a climatological simulation of the North Atlantic ocean circulation, Dyn. Atmos. Oceans, 32, 331-371, 2000.

Meinen, C.S., M.J. McPhaden, and G.C. Johnson, Vertical velocities and transport in the equatorial Pacific during 1993-99, J. Phys. Oceanogr., 31, 3230-3248, 2001.

Perigaud, C., S.E. Zebiak, F. Melin, J-P. Boulanger, and B. Dewitte, On the role of meridional wind anomalies in a coupled model of ENSO,J. Clim., 10, 761-773, 1997.

Reverdin, G. and M.J. McPhaden, Near surface current and temperature variability observed in the equatorial Atlantic from drifting buoys, J. Geoph. Res., 91, 6569-6581, 1986.

Reynolds, R.W. and T.M. Smith, Improved global sea surface temperature analyses using optimum interpolation, J. Clim., 7, 929-948.

Richardson. P.L. and G. Reverdin, Seasonal cycle of velocity in the Atlantic North Equatorial Counter Current as measured by surface drifters, current meters, and ship drift, J. Geoph. Res., 92, 3691-3708, 1987.

Richardson, P.L. and D. Walsh, Mapping climatological seasonal variation of surface currents in the tropical Atalntic using ship drift, J. Geoph. Res., 91, 10,537-10,550, 1986.

Roemmich, D.H., The balance of geostrophic and Ekman transports in the tropical Atlantic, J. Phys. Oceanogr., 13, 1534-1539, 1983.

Swenson, M.S. and D.V. Hansen, Tropical Pacific ocean mixed layer heat budget: The Pacific cold tongue, J. Phys. Oceanogr., 29, 69-81, 1999.

Weisberg, R.H. and L. Qiao, Equatorial upwelling in the central Pacific estimated from moored velocity profilers, J. Phys. Oceanogr., 30, 105-124, 2000.

Zhang, D., M.J. McPhaden, and W.E. Johns, Observational evidence for flow between the subtropical and tropical Atlantic: The subtropical cells, J. Phys. Oceanogr., 2002, submitted.