Search

Loop Current Frontal Eddies (LCFEs) are known to intensify and assist in the Loop Current (LC) eddy shedding. These eddies can also modify the circulation in the eastern Gulf of Mexico (GoM) by attracting water and passive tracers such as chlorophyll and pollutants to the LC-LCFE front. During the 2010 Deepwater Horizon oil spill, part of the oil was entrained not only in the LC-LCFE front but also inside an LCFE, where it remained for weeks. This study assesses the ability of the LCFEs to transport water and passive tracers without exchange with the exterior (i.e., Lagrangian coherence) using altimetry and a high-resolution model. The following open questions are answered: (1) How long can the LCFEs remain Lagrangian coherent at and below the surface? (2) What is the source of water for the formation of LCFEs? (3) Can the formation of LCFEs attract shelf water? The results show that LCFEs are composed of waters originating from the outer band of the LC front, the region north of the LC, and the shelf, and potentially drive cross-shelf exchange of particles, water properties, and nutrients. At depth (~180 m), most LCFE water comes from the outer band of the LC front in the form of smaller frontal eddies. Once formed, LCFEs can transport water and passive tracers in their interior without exchange with the exterior for weeks: these eddies remained Lagrangian coherent for up to 25 days in the altimetry dataset and 18 days at the surface and 29 days at depth (~180 m) in the simulation. LCFE can remain Lagrangian coherent up to a depth of ~560 m. Additional analyses confirm that the LCFE involved in the oil spill formed from water near the oil rig location. Temperature-salinity diagrams show that LCFEs are composed of GoM water as opposed to LC water. Thus, LCFE formation modify the surrounding circulation and the transport of oil and other passive tracers in the eastern GoM.

Continuous measurements from the OSNAP (Overturning in the Subpolar North Atlantic Program) array yield the first estimates of trans-basin heat and salinity transports in the subpolar latitudes. For the period from August 2014 to May 2018, there is a poleward heat transport of 0.50 ± 0.05 PW and a poleward salinity transport of 12.5 ± 1.0 Sv across the OSNAP section. Based on the mass and salt budget analyses, we estimate that a surface freshwater input of 0.36 ± 0.05 Sv over the broad subpolar-Arctic region is needed to balance the ocean salinity change created by the OSNAP transports. The overturning circulation is largely responsible for setting these heat and salinity transports (and the derived surface freshwater input) derived from the OSNAP array, while the gyre (isopycnal) circulation contributes to a lesser, but still significant, extent. Despite its relatively weak overturning and heat transport, the Labrador Sea is a strong contributor to salinity and freshwater changes in the subpolar region. Combined with trans-basin transport estimates at other locations, we provide new estimates for the time-mean surface heat and freshwater divergences over a wide domain of the Arctic-North Atlantic region to the north and south of the OSNAP line. Furthermore, we estimate the total heat and freshwater exchanges across the surface area of the extratropical North Atlantic between the OSNAP and the RAPID-MOCHA (RAPID Meridional Overturning Circulation and Heat-flux Array) arrays, by combining the cross-sectional transports with vertically-integrated ocean heat and salinity content. Comparisons with the air-sea heat and freshwater fluxes from atmospheric reanalysis products show an overall consistency, yet with notable differences in the magnitudes during the observation time period.

Ocean upwelling rates are difficult to measure because of the relatively small velocities involved, and therefore are typically inferred from indirect methods such as heat budget estimates or tracer observations. Here we present the first results using a novel technique, based on the isotope 7 Be, to infer rates of upwelling along the equator. Beryllium-7 (half-life=53.3 d) is a cosmic-ray produced radioactive nuclide that is deposited by rainfall upon the ocean surface and subsequently enriched and homogenized within the mixed layer. Previous investigations have utilized the penetration of characteristically high mixed layer concentrations into the upper thermocline to trace ocean ventilation and subduction over seasonal timescales. Here, the tracer is used in a reverse sense; that is, the 7 Be concentration in the usually 7 Be-rich surface mixed layer will be diluted from penetration of 7 Be “dead” water upwelled from below. This dilution provides a means to infer upwelling rates. Furthermore, with knowledge of upwelling rates, 7 Be profiles can be used to constrain vertical diffusivity within the upper thermocline. These ideas were tested with samples collected during the Tropical Atlantic Climate Experiment (TACE) cruise (May 22–June 27, 2009). The observations indicated a nearly linear relationship between 7 Be inventory and mixed layer temperature, as with increased upwelling, lower mixed layer temperatures correspond to greater 7 Be dilution from depth. With this data, upwelling rates were estimated at a number of stations near the equator between 0°E and 30°W within and adjacent to the equatorial cold tongue. The derived upwelling rates ranged from 0 to 2.2 m/d, with maximum values found between the equator and 2°S. The corresponding K z values derived for the upper thermocline were in the range 1–4×10 −4 m 2 /s.

Efficient monitoring of large-scale current systems for climate research requires the development of new techniques to estimate ocean transports. Here, a methodology for continuous estimation of dynamic height profiles and geostrophic currents from moored temperature sensors is presented. The technique is applied to moorings deployed in the Atlantic Deep Western Boundary Current at 26.5°N, off Abaco, the Bahamas (WOCE ACM-1 array). Relative geostrophic currents are referenced using bottom pressure sensors and available shipboard direct velocity (lowered-ADCP) sections over the period of the deployment, to obtain a time series of absolute volume transport. Comparison with direct velocity measurements from a complete array of current meters shows good agreement for the mean transport and its variablity on time scales longer than 10 days, but larger variability in the current meter derived transport at time scales shorter than 10 days. A rigorous error analysis assesses the contributions of various error sources in the geostrophic as well as direct transport estimates. Low-frequency drift of the bottom pressure sensors is found to be the largest error source in the geostrophic transport estimates and recommendations for improvement of the technique and related measurement technologies are made.

Hydrography is combined with 1-year-long Inverted Echo Sounder (IES) travel-time records and bottom pressure observations to estimate the Deep Western BoundaryCurrent (DWBC) transport east of Abaco Island, the Bahamas (near 26.51N); comparison of the results to a more traditional line of current meter moorings demonstrates that the IESs and pressure gauges, combined with hydrography, can accurately monitor the DWBC transport to within the accuracyof the current meter arrayestimate at this location. Between 800 and 4800 dbar, bounded bytwo IES moorings 82 km apart, the enclosed portion of the DWBC is shown to have a mean southward transport of about 25 Sv (1 Sv ¼ 10 6 m 3 s � 1 ) and a standard deviation of 23 Sv. The DWBC transport is primarilybarotropic (where barotropic is defined as the near-bottom velocityrather than the vertical average velocity ); geostrophic transports relative to an assumed level of no motion do not accuratelyreflect the actual absolute transport variability(correlation coefficient is 0.30). The IES-pressure gauge absolute transport within 1200–4800 dbar agrees well with the current meter absolute transports (upper integration limit based on shallowest current meter level); the standard deviation of the difference is 12 Sv and the mean difference is 0.2 Sv. The correlation coefficient between these two time series is 0.76. The northward flowing Antilles Current (AC) east of Abaco Island has a mean baroclinic transport of 6 Sv as estimated bythe IESs and a standard deviation of 3 Sv. The AC variations observed during 1996–1997 are uncorrelated with the Florida Current transport variations west of Abaco Island in the Florida Straits, however, the AC transport variations bear some resemblance to the historical estimates of the AC annual cycle. r 2004 Elsevier Ltd. All rights reserved.

The Atlantic Meridional Overturning Circulation (AMOC) plays a key role in the global climate system through its redistribution of heat. Changes in the AMOC have been associated with large fluctuations in the earth’s climate in the past and projections of AMOC decline in the future due to climate change motivate the continuous monitoring of the circulation. Since 2004, the RAPID monitoring array has been providing continuous estimates of the AMOC and associated heat transport at 26°N in the North Atlantic. We describe how these measurements are made including the sampling strategy, the accuracies of parameters measured and the calculation of the AMOC. The strength of the AMOC and meridional heat transport are estimated as 17.2 Sv and 1.22 PW respectively from April 2004 to October 2012. The accuracy of ten day (annual) transports is 1.5 Sv (0.9 Sv). Improvements to the estimation of the transport above the shallowest instruments and deepest transports (including Antarctic Bottom Water), and the use of the new equation of state for seawater have reduced the estimated strength of the AMOC by 0.6 Sv relative to previous publications. As new basinwide AMOC monitoring projects begin in the South Atlantic and sub-polar North Atlantic, we present this thorough review of the methods and measurements of the original AMOC monitoring array.

New observations are summarized that lead to the first comprehensive description of the mean inflow distribution in the passages connecting the Atlantic Ocean with the Caribbean Sea. The total Caribbean inflow of 28 Sv is shown to be partitioned approximately equally between the Windward Islands Passages (B10 Sv), Leeward Islands Passages ðB 8S vÞ; and the Greater Antilles Passages ðB10 SvÞ: These results are compared to a numerical model study using a 6-layer, 1=41 resolution Atlantic Basin version of the NRL Layered Ocean Model. Results from two simulations are described, including a purely wind-forced model driven by Hellerman and Rosenstein (J. Phys. Oceanogr. 13 (1983) 1093) monthly winds, and a model with an additional 14 Sv meridional overturning cell driven by inflow/outflow ports at the northern ð651N) and southern ð201S) model boundaries. The purely wind-driven version of the model exhibits a total Caribbean inflow of 17 Sv; consistent with expectations from steady, non-topographic Sverdrup theory. Nearly all of the wind-driven inflow occurs north of Martinique at latitude B151N. The net transport through the Lesser Antilles passages south of 151N (Grenada, St. Vincent, and St. Lucia passages) is nearly zero when the model is forced by winds alone. The addition of a 14 Sv meridional cell in the model increases the net Caribbean inflow to 28 Sv; with nearly all of the additional 11 Sv of inflow entering through the southern Lesser Antilles passages. The modeled inflow distribution resulting from the combined wind and overturning forced experiment is found to compare favorably with the observations. The seasonal cycle of the total inflow in the combined forcing experiment has a mixed annual/semiannual character with maximum in spring and summer and minimum in fall, with a total range of about 4 Sv: The seasonal cycle of the Florida Current resulting from this inflow variation is in good qualitative agreement with observations. Most of the seasonal inflow variation occurs through the Windward Islands passages in the far southern Caribbean, whose annual cycle slightly leads that of the Florida and Yucatan Currents. Variability of the modeled inflow on shorter time scales shows a dramatic change in character moving northward along the Antilles arc. The southern passages exhibit large fluctuations on 30–80 day time scales, which decay to very small amplitudes north of Dominica. Much of this variability is caused by North Brazil Current Rings that propagate northwestward from the equatorial Atlantic and interact with the abrupt island arc topography. The total range of transport variability in individual passages predicted by the model is consistent with observations. However, observations are presently too limited to confirm the seasonal cycles or variability spectra in the Caribbean passages. r 2002 Elsevier Science Ltd. All rights reserved.

The deepest connection between the eastern Caribbean and the Atlantic is over the Jungfern-Grappler Sill complex at 1815 m depth. Through these gaps flows the sole source of water for the deep Caribbean, presumably balanced by diffusively driven upwelling elsewhere in the basin. Fourteen-month-long moored observations at the sills in 1991-1992 reveal a mean transport of 0.11+0.05 Sv (1 Sv -- 106 m3 s-l) of Atlantic water colder than 0=-3.965oC flowing into the Caribbean, one quarter of which comes over the previously unmeasured Grappler Sill. This is about twice the transport seen in previous, shorter experiments. The overflow is highly episodic, with -10 "events" per year. The range of overflow density is comparable to the total vertical strati- fication of the deep Caribbean. A numerical stream-tube model of the overflow plume is run with 13 different initial conditions, representing the observed range of overflow strength and density. Results indicate that much of the overflow only penetrates to 2200-3000 m depth and was dense enough to descend to the bottom of the Caribbean only -1% of the time. Combined plume model results are used to drive a model of the basin stratification. A steady balance is found to be possible with small diapycnal diffusivity in the basin, between 0.1 and 0.6 cm2 s-1. The reason for the small diffusivity is that the stratification is essentially built in place by the different detrainment depths, and so less diffusive transformation is needed. Below a depth of 1815 m the eastern Caribbean is topog- raphically isolated from the rest of the worlds' ocean (Figure l a). Middepth water flows from the Atlantic through Anegada Pass (1915 m) into the Virgin Islands Basin (VIB) (Figure lb), a small, intermediate basin between the Atlantic and Caribbean. The source water for the inflow is Upper North Atlantic Deep Water (UNADW) which lies between depths of 1200-2000 m along the western boundary of the North Atlantic. Leaman and Harris (1990) show the Freon maximum at 26.5oN associated with Labrador Sea Water (nominally the source of UNADW)to be at 1200 m, considerably shallower than the Anegada sill depth; however, Labrador Sea Water still appears to form the bulk of what flows into the deep Caribbean. Atlantic water at that depth is also influenced to a lesser degree by the high-salinity Mediterranean Ouflow Water, as is apparent from the contour map of salinity on the 3.8oC potential temperature surface by Worthington and Wright (1970). From the VIB, water flows through Jungfern Pass (1815 m) and nearby Grappler Pass (1710 m) as reported by Fratantoni et al. (1997), forming a dense overflow plume that cascades into the deep Caribbean. (Throughout this paper

During 1991–1994, repeated measurements of current structure and water mass properties were made in the major southern passages to the Caribbean Sea between Trinidad and Dominica. A total of 10 cruises were performed in different seasons, consisting primarily of serial station occupations in the Grenada, St Vincent, St Lucia, and Dominica Passages. This data set is by far the most comprehensive available in these passages and better determines the mean flow and range of variability than do previous studies. The flow structure in these passages is characterized by a strong and relatively stable inflow above the thermocline (approximately the upper 100 m), and a more highly variable flow regime within and below the thermocline. Typical near-surface inflow velocities in Grenada and St Vincent passages were 40–60 cm s −1 , with maximum observed currents of 90 cm s −1 . Frequently, counterflows were observed below this surface layer flowing out of the Caribbean, trapped to the southern side of the passages. This subsurface counterflow appeared to be strongest and most prevalent in Grenada Passage, where outflow speeds as large as 30 cm s −1 were observed. As a result, the ensemble mean sections for the three southern passages exhibit a deep inflow concentrated in the northern and central parts of the passages, and weak outflow on the southern sides. The ensemble of measurements from this program indicate a mean transport of 9.5±3 × 10 6 m 3 s −1 through the southern passages, with a range of variability from 3 to 17 × 10 6 m 3 s −1 . No clear annual cycle is apparent in the passage transports. Mean transports calculated for the individual passages were 4.7, 3.4, 0.9, and 0.5 × 10 6 m 3 s −1 , respectively, for Grenada, St Vincent, St Lucia, and Dominica Passages, decreasing monotonically northward and indicating that Grenada and St Vincent Passages largely dominate the inflow to the southern Caribbean. Consideration of this data set, historical data, Sverdrup theory and numerical model results suggests that very little of the mean inflow to the Caribbean through the southern passages can be attributed to wind-driven circulation, and that most of the transport (6–8 × 10 6 m 3 s −1 ) is derived from South Atlantic waters crossing the equator in the upper limb of the Atlantic thermohaline cell. The remainder of the upper ocean thermohaline transport likely follows a different route into the subtropical gyre, possibly flowing northward in the interior and entering the Caribbean through the more northern passages.

The Amazon shelf is subject to energetic forcing from a number of different sources, including near-resonant semi-diurnal tides, large buoyancy flux from the Amazon River discharge, wind stress from the northeasterly tradewinds and strong along-shelf flow associated with the North Brazil Current. Although the large volume of river discharge produces a pronounced salinity anomaly, the water motions on the shelf are dominated by the other forcing factors. Tidal velocities of up to 200 cm s−1 are generally oriented in the cross-shelf direction. Tide-induced mixing influences the position and structure of the bottom salinity front that separates the well-mixed nearshore region from the stratified plume. High concentrations of suspended sediment trapped along the frontal zone increase the stability of the tidal boundary layer and thus reduce the bottom stress. At subtidal frequencies, motion is primarily along-shelf toward the northwest, both in the plume and in the ambient, high-salinity water of the outer-shelf. The plume is generally 5–10 m thick, with a salinity of 20–30 psu. The along-shelf velocity within the plume varies as a function of the along-shelf wind stress. This variability results in large temporal variations in plume structure and freshwater content on the shelf. The net northwestward motion of the Amazon plume and of the ambient shelf water appears to be the result of a large-scale pressure gradient associated with the North Brazil Current system.

[1] As part of the global thermohaline circulation, some North Atlantic Deep Water (NADW) exits the Atlantic basin to the south of Africa. Observations have shown that there is a quasi-zonal pathway centered at 25°S carrying NADW eastward, connecting the Deep Western Boundary Current to the Cape Basin. However, it has been unclear what sets this pathway. In particular, waters must move southward through the Cape Basin, thereby crossing isolines of planetary vorticity, in order to exit the basin. Here, we find that an eddy thickness flux induced by Agulhas rings moving northwestward forces a circulation of NADW through the Cape Basin. The pathway at 25°S feeds the southeastward flow of this circulation while conserving potential vorticity. Using Lagrangian floats advected for 300 years in a 1/10° resolution ocean model, we show that the most common pathway for NADW in our model lies directly below the Agulhas ring corridor. By analyzing the velocity and density fields in the model, we find that the decay of these rings, and their forward tilt with depth, results in a southward velocity, across isolines of planetary vorticity, of 1 to 2 cm/s in the deep waters. The associated stream function pattern yields a deep circulation transporting 4 Sv of NADW from the Deep Western Boundary Current at 25°S to the southern tip of Africa.

[1] Observations are used to develop metrics for interannual tropical instability wave (TIW) variability in the Atlantic and to relate that variability to larger scale processes. The analysis is partitioned into different latitude bands to distinguish between off-equatorial (5°S, 2°N, and 5°N) and near-equatorial (2°S and 0°) TIWs. TIW metrics based on sea surface temperature (SST) and sea level anomaly (SLA) fluctuations are compared against interannual anomalies of SST in the cold tongue region. To examine the role of barotropic shear instabilities in modulating the intensity of a TIW season, wind stress and near-surface current indices are developed in regions where the shear between the Equatorial Undercurrent (EUC) and the northern branch of the South Equatorial Current (nSEC) and between the nSEC and the North Equatorial Countercurrent (NECC) are expected to be largest. Good agreement is found between the SST and SLA TIW metrics along the off-equatorial latitude bands, and interannual variations of both metrics can largely be attributed to barotropic shear instabilities. In particular, years with low (high) TIW variance along the off-equatorial latitude bands are associated with anomalously warm (cold) SSTs in the cold tongue region, weak (strong) wind stress divergence and curl in the EUC-nSEC region, and weak (strong) zonal current shear in the nSEC-NECC region. In contrast, in the near-equatorial latitude bands, poor agreement is found between interannual TIW activity based on the SST and SLA metrics, and near-equatorial TIW variability cannot be explained by the large-scale SST, wind stress divergence and curl, and current shear indices.

More than two decades of hydrography on the Abaco line east of the Bahamas at 26 degrees N reveals decadal variability in the salinity of classical Labrador Sea Water (cLSW), despite the long distance from its source region in the North Atlantic Ocean. Hydrographic time series from the Labrador Sea and from the Abaco line show a pronounced step-like decrease in salinity between 1985 and 1995 in the Labrador Sea and between 1995 and 2010 at the Abaco line, suggesting a time lag between the two locations of approximately 9 years. The amplitude of the anomaly at the Abaco line is 50% of the amplitude in the Labrador Sea. A similar time lag and reduction of amplitude is found in the high-resolution OFES model, in which salinity anomalies can be observed propagating through the Deep Western Boundary Current as well as through a broad interior pathway. On its way south to the Abaco line, the cLSW becomes 8 standard deviations saltier due to isopycnal mixing with Mediterranean Outflow Water (MOW). Climatological data in the North Atlantic suggests that the mixing ratio of MOW to cLSW at the Abaco line is 1:4 and that no variability in MOW is required to explain the observed variability at the Abaco line. The data studied here suggest that decadal cLSW anomalies stay relatively coherent while getting advected, despite the important role of interior pathways.

[1] In the western tropical Atlantic, the North Brazil Current retroflection periodically sheds large anticyclonic rings, which then propagate northwestward. Between 1998 and 2000, the North Brazil Current Rings Experiment sampled a large number of these rings by shipboard and moored acoustic Doppler current profiler. Ten of the sampled rings are analyzed in this study, focusing on their sea surface dynamic properties. The rings exhibit a radial structure consisting of two regimes, an “inner” core region in near solid body rotation and an “outer” ring regime with an approximately exponentially decaying structure. The observations show a sharp change in vorticity at the regime transition and the presence of a strong opposite vorticity shield bounding the inner solid body core. We show that Gaussian models, commonly used to represent the surface expression of these and other rings, are adequate for determining the sea surface height anomaly but tend to poorly estimate other properties such as the maximum swirl velocity. Therefore, we propose a new two-part model as a better approximation of the rings' radial structure. According to the cyclogeostrophic balance approximation, the sea surface height distribution across the inner ring has a parabolic shape, while the outer ring has an exponential structure similar to the velocity field. Interestingly, many of the observed rings have an intensity very close to the theoretical limit for anticyclones at these latitudes, which is believed to be due to inertial instability.

Observations from a year-long deep current meter mooring at 8°N, 52°W are used to describe the structure and variability of the Deep Western Boundary Current (DWBC) in the tropical Atlantic. The DWBC has a deep core near 4300 m depth, is extremely swift and narrow (35 cm s −1 core speed, 60 km width), and shows a total transport of 22 Sv between 2500 m and the bottom. Approximately 3 Sv of the DWBC transport are carried in near-bottom layers below θ =1.8° C , which appears to be recirculating in the western tropical Atlantic north of about 4°N. Variations in the DWBC are dominated by large-amplitude vertical displacements of the deep temperature field at 60–70 day periodicities that cause significant changes in the deep stratification and vertical transport structure of the DWBC. These fluctuaions appear to be related to periodic surges of cold bottom waters up onto the base of the continental rise that become entrained in the DWBC. While the forcing mechanism for these deep fluctuations remains unexplained, the picture emerging from these data is that of a highly active abyssal layer in the tropical Atlantic.

[1] The circulation and water mass transformation processes in the Persian Gulf and the water exchange with the Indian Ocean through the Strait of Hormuz are studied using the Hybrid Coordinate Ocean Model (HYCOM). Model results show that the Indian Ocean Surface Water (IOSW) propagates in two branches into the gulf, one along the Iranian coast toward the northern gulf and the other one onto the southern banks driven by the Ekman drift due to the prevailing northwesterly winds. These two branches of inflow form two cyclonic gyres in the northern and in the southern gulf. A salinity front separates the fresher intruding IOSW from the saltier waters in the gulf. Eddies with size of about 100 km are fully developed along the salinity front in summer. The intrusion of the IOSW in the model extends much farther into the gulf in summer than in winter. By analyzing the salt balance in the basin and conducting sensitivity experiments, we show that it is the balance between the advection of IOSW and vertical salt flux induced by mixing that mainly controls the seasonal variation of the surface salinity. Surface wind stress plays a secondary role in modulating the seasonal intrusion of the IOSW. High-frequency atmospheric forcing produces more realistic surface temperatures than obtained from climatological forcing, as a result of increased heat loss in winter. However, the high-frequency forcing does not change significantly the general features of the circulation.

[1] The circulation and water mass transformation processes in the Persian Gulf and the water exchange with the Indian Ocean through the Strait of Hormuz are studied using the Hybrid Coordinate Ocean Model (HYCOM). Model results suggest that intruding Indian Ocean Surface Water (IOSW) is transformed into hypersaline waters with salinity >41 practical salinity unit by the fresh water loss in the northern end of the gulf and in the southern shallow banks. During wintertime, intense heat loss from sea to air leads to the formation of cold and saline waters both in the northern gulf and in the southern gulf. Dense waters formed in the southern gulf have higher salinity and spill into the axial deep trough of the gulf as a sporadic bottom outflow in winter, which leads to high-salinity pulses in the strait as observed, whereas in summer their buoyancy is increased due to heating and they are exported as a warm yet salty intermediate depth flow through the strait. Dense waters formed in the northern gulf propagate toward the strait along the axial trough throughout the year. Correspondingly, there are two branches of seasonal overturning circulation in density classes: the northern branch peaking in August with strength of 0.13 Sv and the southern branch peaking in February and December with strength of 0.08 Sv. These two branches outflow together feed a seasonally varying deep outflow through the strait with an annual mean volume transport of 0.12 Sv.

[1] A dual-station high-frequency Wellen radar (WERA), transmitting at 16.045 MHz, was deployed along the eastern Florida Shelf and operated and maintained by the University of Miami's Rosenstiel School of Marine and Atmospheric Science. From September 2004 to June 2005, a moored acoustic Doppler current profiler (ADCP) acquired subsurface current measurements within the radar footprint along the shelf break at 86-m depth. The shallowest ADCP bin located at 14-m depth was used as a comparison for the WERA surface measurements. RMS differences ranged from 0.1 to 0.3 m s -1 between the surface and 14-m depth, with good agreement over most of the period. Regression analyses indicated slopes near unity in the north-south (v) component and ≈0.5 for the east-west (u) component velocities. When utilized in tandem with the ADCP subsurface measurements, WERA enables three-dimensional snapshots of coastal oceanographic features to be resolved. For example, from December 2004 through February 2005, three energetic circulation patterns were observed: (1) a subsurface stratified countercurrent, (2) a submesoscale coherent vortex, and (3) a mesoscale circulation feature, i.e., a propagating Tortugas gyre. These features represent the significant current variability along the western flank of the Florida Current that impacts the coastal ocean.

[1] The Agulhas Undercurrent Experiment took place in February–March 2003 off the east coast of South Africa and consisted of four sections of hydrographic and velocity data across the Agulhas Current between 30 and 36°S and connecting offshore sections that formed three closed boxes. An inverse model was applied to the quasi-synoptic data, and results show that the net mass transport at the historical 32°S section had a considerably higher transport of 100 ± 9 Sv than earlier estimates. This high transport falls within the peak-to-peak variability obtained previously from a current meter time series. Several mesoscale cyclonic eddies extending down to intermediate depths were sampled during the survey; in particular, a strong, locally formed shear edge eddy was found inshore of the Agulhas Current at 36°S. Offshore eddies were found to drive considerable onshore-offshore fluxes, resulting in highly variable Agulhas transports from one section to another. After attempting to account for and remove the influence of these eddies on the Agulhas transport, the downstream growth of the Agulhas Current is found to be consistent with the Sverdrup transport variation, within errors. To account for the total magnitude of the Agulhas transport, fluxes from both the Indonesian Throughflow and Indian Ocean overturning must also be taken into account. There is no clear evidence from this study for a significant contribution of inertial recirculation to the Agulhas within this latitude range.

[1] Aiming at exploring and understanding the summer circulation in the Red Sea, a cruise was conducted in the basin during the summer of 2001 involving hydrographic, meteorological, and direct current observations. The most prominent feature, characteristic of the summer circulation and exchange with the Indian Ocean, is a temperature, salinity, and oxygen minimum located around a depth of 75 m at the southern end of the basin, associated with Gulf of Aden Intermediate Water inflowing from the Gulf of Aden during the summer season as an intruding subsurface layer. Stirring and mixing with ambient waters lead to marked increases in temperature (from 16.5 to almost 33°C) and salinity (from 35.7 to more than 38 psu) in this layer by the time it reaches midbasin. The observed circulation presents a very vigorous pattern with strong variability and intense features that extend the width of the basin. A permanent cyclone, detected in the northern Red Sea, verifies previous observations and modeling studies, while in the central sector of the basin a series of very strong anticyclones were observed with maximum velocities exceeding 1 m/s. The three-layer flow pattern, representative of the summer exchange between the Red Sea and the Gulf of Aden, is observed in the strait of Bab el Mandeb. In the southern part of the basin the layer flow is characterized by strong banking of the inflows and outflows against the coasts. Both surface and intermediate water masses involved in the summer Red Sea circulation present prominent spatial variability in their characteristics, indicating that the eddy field and mixing processes play an important role in the summer Red Sea circulation.

[1] New satellite-based observations reveal that westward translating anticyclonic rings are generated as a portion of the Somali Current accelerates northward through the Socotra Passage near the mouth of the Gulf of Aden. Rings thus formed exhibit azimuthal geostrophic velocities exceeding 50 cm/s, are comparable in overall diameter to the width of the Gulf of Aden (250 km), and translate westward into the gulf at 5-8 cm/s. Ring generation is most notable in satellite ocean color imagery in November immediately following the transition between southwest (boreal summer) and northeast (winter) monsoon regimes. The observed rings contain anomalous fluid within their core which reflects their origin in the equator-crossing Somali Current system. Estimates of Socotra Passage flow variability derived from satellite altimetry provide evidence for a similar ring generation process in May following the winter-to-summer monsoon transition. Cyclonic recirculation eddies are observed to spin up on the eastern flank of newly formed rings with the resulting vortex pair translating westward together. Recent shipboard and Lagrangian observations indicate that vortices of both sign have substantial vertical extent and may dominate the lateral circulation at all depths in the eastern Gulf of Aden.

The exchange between the Persian (Arabian) Gulf and the Indian Ocean is investigated using hydrographic and moored acoustic Doppler current profiler data from the Straits of Hormuz during the period December 1996 to March 1998. The moored time series records show a relatively steady deep outflow through the strait from 40 m to the bottom with a mean speed of approximately 20 cm/s. A variable flow is found in the upper layer with frequent reversals on timescales of several days to weeks. The annual mean flow in the near-surface layer is found to be northeastward (out of the Persian Gulf) in the southern part of the strait, suggesting a mean horizontal exchange with the Indian Ocean that is superimposed on the vertical overturning exchange driven by evaporation over the gulf. The salinity of the deep outflow varies from 39.3 to 40.8 psu with highest outflow salinities occurring in the winter months (December–March). The annual mean deep outflow through the strait is estimated to be 0.15 ± 0.03 Sv. Calculation of the associated heat and freshwater fluxes through the strait yields estimates for the annual heat loss over the surface of the gulf of ?7 ± 4 W/m2 and an annual water loss (E-P-R) of 1.68 ± 0.39 m/yr. These values are shown to be in relatively good agreement with climatological surface fluxes derived from the Southampton Oceanography Centre global flux climatology after known regional biases in the radiative budget are taken into account.

[1] The three-dimensional circulation of the Red Sea is studied using a set of Miami Isopycnic Coordinate Ocean Model (MICOM) simulations. The model performance is tested against the few available observations in the basin and shows generally good agreement with the main observed features of the circulation. The main findings of this analysis include an intensification of the along-axis flow toward the coasts, with a transition from western intensified boundary flow in the south to eastern intensified flow in the north, and a series of strong seasonal or permanent eddy-like features. Model experiments conducted with different forcing fields (wind-stress forcing only, surface buoyancy forcing only, or both forcings combined) showed that the circulation produced by the buoyancy forcing is stronger overall and dominates the wind-driven part of the circulation. The main circulation pattern is related to the seasonal buoyancy flux (mostly due to the evaporation), which causes the density to increase northward in the basin and produces a northward surface pressure gradient associated with the downward sloping of the sea surface. The response of the eastern boundary to the associated mean crossbasin geostrophic current depends on the stratification and b-effect. In the northern part of the basin this results in an eastward intensification of the northward surface flow associated with the presence of Kelvin waves while in the south the traditional westward intensification due to Rossby waves takes place. The most prominent gyre circulation pattern occurs in the north where a permanent cyclonic gyre is present that is involved in the formation of Red Sea Outflow Water (RSOW). Beneath the surface boundary currents are similarly intensified southward undercurrents that carry the RSOW to the sill to flow out of the basin into the Indian Ocean. INDEX TERMS: 4243 Oceanography: General: Marginal and semienclosed seas; 4255 Oceanography: General: Numerical modeling; 4532 Oceanography: Physical: General circulation; KEYWORDS: Red Sea, marginal sea, Oceanic General Circulation Model, water mass formation

[1] The mechanisms involved in the seasonal exchange between the Red Sea and the Indian Ocean are studied using an Oceanic General Circulation Model (OGCM), namely the Miami Isopycnic Coordinate Ocean Model (MICOM). The model reproduces the basic characteristics of the seasonal circulation observed in the area of the strait of Bab el Mandeb. There is good agreement between model results and available observations on the strength of the exchange and the characteristics of the water masses involved, as well as the seasonal flow pattern. During winter, this flow consists of a typical inverse estuarine circulation, while during summer, the surface flow reverses, there is an intermediate inflow of relatively cold and fresh water, and the hypersaline outflow at the bottom of the strait is significantly reduced. Additional experiments with different atmospheric forcing (seasonal winds, seasonal thermohaline air–sea fluxes, or combinations) were performed in order to assess the role of the atmospheric forcing fields in the exchange flow at Bab el Mandeb. The results of both the wind- and thermohaline-driven experiments exhibit a strong seasonality at the area of the strait, which is in phase with the observations. However, it is the combination of both the seasonal pattern of the wind stress and the seasonal thermohaline forcing that can reproduce the observed seasonal variability at the strait. The importance of the seasonal cycle of the thermohaline forcing on the exchange flow pattern is also emphasized by these results. In the experiment where the thermohaline forcing is represented by its annual mean, the strength of the exchange is reduced almost by half.

Moored observations of Kuroshio current structure and transport variability were made across the channel between northeast Taiwan and the Ryukyu Islands at 24 degreesN from September 19, 1994, to May 27, 1996. This was a cooperative, effort between the United States and Taiwan. The moored array was designated PCM-1, for the World Ocean Circulation Experiment (WOCE) transport resolving array. The dominant current and transport variability occurred on 100-day timescales and is shown by Zhang et al. [2001] to be caused by warm mesoscale eddys merging with the Kuroshio south of the array causing offshore meandering and flow splitting around the Ryukyu Islands. An annual transport cycle could not be resolved from our 20-month moored record because of abasing from the 100-day period events. Sea level difference data were used to extend the transport time series to 7 years giving a variation in the range of the annual transport cycle of 4-10 Sv, with a mean range closer to 4 Sv. The seasonal maximum of 24 Sv occurred in the summer and the seasonal minimum of 20 Sv occurred in the fall. A weaker secondary maximum also occurred in the winter. The cycle of Kuroshio transport appears to result from a combination of local along-channel wind forcing and Sverdrup forcing over the Philippine Sea. Our estimate of the mean transport of the Kuroshio at the entrance to the East China Sea from the moored array is 21.5 +/- 2.5 Sv. The mean transpacific balance of meridional flows forced by winds and thermohaline processes at this latitude requires an additional mean northward flow of 12 Sv with an annual cycle of +/-8 Sv along the eastern boundary of the Ryukyu Islands. The mean transport and annual cycle of the Kuroshio were found to be in reasonable agreement with basin-scale wind-forced models. Remarkable similarities are shown to exist between the mean western boundary currents and their seasonal cycles in the Atlantic (Florida Current and Antilles Current) and Pacific (Kuroshio and boundary current east of Ryukyu Island chain) at the same latitude. However, detailed comparison shows that the mean Kuroshio is weaker and more surface intensified than the mean Florida Current, while the Kuroshio-transport variability is significantly larger.

Erratum to ‘‘On the Atlantic inflow to the Caribbean Sea’’ [Deep-Sea Research I 49 (2002) 211–243] William E. Johns*, Tamara L. Townsend, David M. Fratantoni, W. Douglas Wilson Rosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149, USA Naval Research Laboratory, Stennis Space Center, MS 39529, USA Department of Physical Oceanography, Woods Hole Oceanographic Institute, Woods Hole, MA 02543, USA Physical Oceanography Division, NOAA/AOML, 4301 Rickenbacker Causeway, Miami, FL 33149, USA

Large anticyclonic rings are shed from the retroflecting North Brazil Current (NBC) near 8 deg N in the tropical western Atlantic. New subsurface velocity and temperature measurements within three such rings are presented here and are found to be consistent with previous in situ and remotely sensed NBC ring measurements. A high-resolution numerical model of the Atlantic Ocean forced by monthly wind stress and an imposed meridional overturning cell is found to shed NBC rings that approximate those observed. The model rings are more surface-intensified than those observed and somewhat smaller in diameter. Both observed and modeled NBC rings move northwestward along the coast of South America with a speed of 8-16 cm/s, considerably slower than predicted by analytical theories describing westward ring propagation. At least 2-3 rings per year separate from the NBC retroflection. Annually, 1-3 rings translate intact from their formation region near 50 deg W to the islands of the southeastern Caribbean, where they disintegrate after a lifetime of about 100 days. This study is the first to make use of subsurface temperature and velocity data to compute the volume of the anomalous ring core. NBC rings may be responsible for 3-4 Sv of direct mass transport across the equatorial-tropical gyre boundary or 20-25% of the total upper ocean cross-gyre transport required by the Atlantic meridional overturning cell. Translating NBC rings may contribute 20% of the total meridional heat transport by the ocean at this latitude.

An analysis of the structure and transport of the Gulf Stream is undertaken using direct current meter observations from a 13-mooring array deployed near 68°W from June 1988 to August 1990. The analysis is based on a “stream-coordinate” approach, in which velocities are rotated into a local, downstream coordinate frame and averaged according to their relative cross-stream location within the current. The picture so obtained represents the average synoptic structure of the Gulf Stream, rather than the Eulerian-averaged structure in which the current is weakened and broadened by lateral meandering of the current and adjacent recirculations. Many familiar features of the Gulf Stream are reproduced in the analysis, including an asymmetric velocity profile with larger shear on the cyclonic (shoreward) side of the current, an offshore displacement of the velocity core with depth, and a subsurface velocity maximum on the offshore side of the current. Westward recirculations are also seen on both sides of the Gulf Stream. Maximum downstream speeds at the axis of the Gulf Stream reach approximately 2.0 m/s at the surface and 0.7 m/s at 1000 m, roughly twice the corresponding Eulerian-averaged values. The analysis also reveals a deep extension of the Gulf Stream at 3500 m depth with a width of 130 km and average speeds of 3–4 cm/s. The transport of the Gulf Stream in the stream-coordinate frame is 113±8 Sv, approximately 30% larger than the Eulerian-averaged transport of 88 Sv. On the basis of these results and other recent studies the downstream transport increase of the Gulf Stream and the inflow structure to the Gulf Stream are reconsidered. It is concluded that approximately 30 Sv, or over half of the transport increase between Cape Hatteras and 68°W, is fed by inflow from the northern side of the Gulf Stream and that this inflow is concentrated near Cape Hatteras and 68°W, where the Gulf Stream flows steeply across isobaths converging from the north.

We report results from a 1-year (September 1987 to September 1988) moored current meter array spanning the continental margin off French Guiana near 8° N in the western tropical Atlantic. Current profiles were recorded at three sites: at the shelf break, over the mid-continental slope, and at the base of the continental rise. Upper level mean currents showed a northwestward flowing North Brazil Current (NBC) and offshore retroflection of this flow into the North Equatorial Countercurrent from late summer through about January. Generally weak upper level mean flows were observed during the spring (February–June). Persistent northwestward mean flow was observed at 900 m depth over the continental slope, indicating northward transport of Antarctic Intermediate Waters in a subsurface boundary flow at speeds of 10–15 cm s−1. Deep currents over the continental rise showed a strong southeastward Deep Western Boundary Current (DWBC) extending from 2500 m to the bottom, with mean core speeds of nearly 30 cm s−1 at 4300 m depth. Transport estimates based on these data and a few geostrophic sections suggest a DWBC transport of 20–40 × 106 m3 s−1 at this location. Low-frequency current fluctuations were dominated by a well-defined 40- to 60-day oscillation with peak-to-peak meridional velocity amplitudes of > 1 m s−1 during the fall. Analysis of historical coastal zone color scanner imagery suggests that these oscillations are related to quasi-periodic generation and subsequent westward movement of ≈ 400 km diameter eddies from the NBC retroflection. These results contrast sharply with earlier indications of a quasi-permanent “Demerara Eddy” in this region, and suggest that this commonly observed feature is in fact a transient phenomenon associated with the time-dependent behavior of the NBC retroflection.

A self-contained, upward-looking acoustic Doppler current profiler (ADCP), mounted in the top float of a subsurface mooring was deployed in the Somali Current at 2°14′N, 45°55′E from September 17, 1985, to April 25, 1986. The instrument operated at a frequency of 150 kHz, with a vertical beam angle of 20°. Vector-averaged profiles of horizontal and vertical velocity were recorded every 4 hours, using 200 pings per ensemble at a vertical bin length of 8.7 m. The mooring was deployed in very rough topography, settling in a trough at 337 m depth with the ADCP located at 267 m depth. Data retrieval over the entire recording period was complete, with Doppler biasing from side lobe reception of vertically traveling rays affecting only the top 20 m below the surface. Over the 7-month deployment the instrument recorded current profiles encompassing the end of the 1985 summer monsoon and entire winter monsoon and also through the spring transition into the early onset phase of the 1986 summer monsoon. Significant echo amplitude variations of week-to-month-long duration were observed, which were only partially related to horizontal flow variations associated with the monsoons. Projection of the strong horizontal currents (exceeding 150 cm/s at times) into the vertical component was not observed, attesting to fairly exact orientation of the four beams and tilt meters. This indicates that the vertical current measurement from ADCPs can be potentially useful for phenomena with vertical velocities exceeding a few millimeters per second. However, an analysis of echo amplitude and vertical current variations at the diurnal period suggests that the measured vertical velocity is, at least at that period, probably dominated by active vertical migration of biological scatterers through the water column.

We present a new method for continuously tracking the location of the Gulf Stream using a moored array of inverted echo sounders. Time series of lateral displacements of the front, shown accurate to -+8 km, have been collected along three sections spaced 100, 150, and 200 km downstream (NE) of Cape Hatteras, North Carolina, for a period of 12 months. These records are highly coherent at all periodicities longer than 4 days. From the observed phase lags, a dispersion relationship is presented for the meanders' As the period and wavelength (T, h) increase from (4 days, 180 km) to (30 days, 600 km), the phase speed decreases smoothly from 40 to 20 km d-. The meanders exhibited rapid growth at periods longer than 4 days, doubling in variance in each 50-km step downstream. This downstream growth is most appropriately described by a spatial e-folding scale K - - 400 km for (T, X) (14 days, 400 km). from an array of seven inverted echo sounders (IES) and three deep current meters deployed 100 to 200 km northeast of Cape Hatteras in 2000-4000 m water. This array continu- ously monitored the path of the Gulf Stream and the deep currents under it. On four deployment and recovery cruises to the area, expendable bathythermograph traces (XBT) and hydrographic data were collected to survey the temperature and density structure along the stream path and on sections normal to it. After a brief review of earlier studies on meanders, this paper presents a discussion of the measurement techniques using IES in the Gulf Stream. We show by direct compari- sons with traditional measurement methods that (1) each IES produces a continuous record of the thermocline depth (here taken to be the 15oC isotherm for subsequent comparison with standard methods of determining the position of the 'north wall' of the stream) with accuracy of 20 m in a range of variation of 800 m, and (2) the array of IES's paired across the stream produces continuous estimates of the stream position with 8 km accuracy. The results from the three IES sections located 100, 150, and 200 km downstream of Cape Hatteras show very high coherence throughout the array with rapid (20 to 40 km d -l) translation of the meanders downstream accompanied by rapid growth of the lateral displacements, doubling in variance with each 50-km step downstream. In a companion paper (W. E. Johns and D. R. Watts, manuscript in preparation, 1982) we discuss the current meter records and the relationship of the deep currents to Gulf Stream meanders. 1981); they have wavelengths of 150-200 km and propagate downstream at rates of approximately 30-40 km d-. Downstream of Cape Hatteras the envelope of meandering broadens greatly to 200-300 km (Figure 1), and the stream can shift through more than its own width. In the early shipboard surveys (Fuglister and Worthington, 1951;Fuglis- ter, 1963) the path of the stream south of Nova Scotia was dominated by long-wavelength (-400 km), large-amplitude meanders. Those which were studied in detail were found to be quasi-stationary and were characterized by a slowly evolving baroclinic field, indicating temporal growth. Far- ther upstream near 72oW, however, significant eastward propagation was observed; one meander crest moved at a rate of approximately 17 km d - for 3 days (Fuglister and Worthington, 1951). Hansen (1970) later conducted monthly path surveys between 75 o and 55oW for 1 year, with weekly surveys over a 2-month period, using a towed thermistor. He found meanders on appreciably shorter scales, with wave- lengths ranging from 200 to 400 km. Mean downstream propagation rates of 5-10 cm s - and spatial growth rates of 2 x 10 -3 km - were estimated by constructing lines of constant phase between features on successive path se- quences. The rapid evolution between monthly paths, how- ever, produced considerable ambiguity in identifying and tracking events. Weekly paths evolved more continuously and were apparently not aliased by higher-frequency/wave number phenomena.