Click here for the graphics an figures associated with this paper
(117 MB PDF)

Introduction

The Dry Tortugas are located at the northern boundary of the southern entrance to the Straits of Florida. This location is unique for it is at the cross-roads between the eastern Gulf of Mexico and the Atlantic and is thereby strongly impacted by major currents and eddy systems that connect these regions. The islands that make up the Dry Tortugas lie adjacent to the paths of both the Loop Current in the eastern Gulf of Mexico and the Florida Current in the Straits of Florida. The marine ecosystems of the Tortugas are therefore subject to foreign inputs of natural and anthropogenic origins from remote sources 100’s to 1000’s km distant by the connective flow that forms these major current systems. Also significant to the ecosystems are the large eddies that form along the inshore boundaries of the current systems and form retention areas that support recruitment of locally spawned larvae. Recruitment pathways are further influenced by the bathymetry of the region, together with wind and tidal currents of the west Florida shelf and Florida Keys, and the seasonal development of shelf stratification. The occurrence of strong transient events is also not uncommon in the Tortugas due to the occasional passage of storms, cold-air outbreaks and intrusions of Mississippi River water.

The intent of this chapter is to synthesize the available information on physical oceanographic processes in the Dry Tortugas using previous works contained in the literature and results from ongoing studies. Also included is discussion of the influence of physical processes on the larval recruitment from local and remote sources.

Background

The Dry Tortugas form the western extent of the Florida Keys island chain (Fig. 1). Water depths in the Tortugas are typically less than 20 m, but the area is surrounded by deeper waters with depths of 40 m or greater to the south, west and north, due to its position at the junction of the Straits of Florida, Gulf of Mexico and west Florida shelf, respectively. The Tortugas are separated from the shallower reef tract of the Florida Keys to the east by Rebecca Channel with depths of 20 to 30 m.

The Straits of Florida are oriented in an east-west direction in the southern region below 25o N, then turn cyclonically to a north-south orientation above this latitude (Fig. 1). Water depths in the Straits decrease from about 2000 m at the westward entrance to less than 800 m in the northern Straits. The southern Straits of Florida (SSF) are bounded by the Florida Keys to the north and Cuba to the south.

Flows in the eastern Gulf of Mexico and SSF are constrained by the configuration of the Straits and broad shallow shelves of the Campeche Bank and the west Florida shelf (Fig. 1). The Yucatan Current emerging from the Yucatan Channel is steered northward into the eastern Gulf of Mexico by the Campeche Bank (Fig. 2). The current then turns anticyclonically to flow southward along the steep escarpment of the west Florida shelf before entering the SSF near the Tortugas. The Straits of Florida form a conduit for the Florida Current (FC), which connects the flow out of the eastern Gulf of Mexico, the Loop Current, with the Gulf Stream in the North Atlantic (Fig. 2). The FC is made up of about equal parts of waters originating in the South Atlantic and North Atlantic subtropical gyres (Schmitz and Richardson, 1991; Wilson and Johns, 1997) and is therefore an important link in both the North Atlantic Sverdrup circulation (Leetmaa et al., 1977) and the global thermohaline circulation (Gordon, 1986). The upper layer waters of the FC with temperatures greater than 24 oC are derived primarily from the South Atlantic (Schmitz and Richardson, 1991), and are transported across the equator and through the Caribbean by the combined influence of the North Brazil Current and the North Atlantic wind-driven subtropical gyre.

A considerable background of information exists on FC variability in the northern Straits, and on the Loop Current in the eastern Gulf of Mexico, but until recently, very little was known about the flow behavior in the southern Straits of Florida. Early observations of flow in the SSF by Brooks and Niiler (1975) found the axis of the FC to be located about 80 km offshore of Key West during a one month period of dropsonde current measurements in the summer of 1972, and a persistent counterflow occurred along the northern shore. They interpreted this counterflow as a cyclonic recirculation of FC water. Chew (1974) also showed evidence in hydrographic and drifter data of a recirculation feature shoreward of a large offshore meander of the FC in the lower Keys. Vukovich (1988) used satellite thermal imagery to conclude that a quasi-persistent cold perturbation exists off the Dry Tortugas, where the southward flowing Loop Current turns eastward into the SSF.

The above early observations of flow variability in the SSF are in marked contrast to the behavior of the FC in the northern Straits and to the Loop Current. The current axis in the northern Straits is located about 25 km offshore on the mean, with no persistent counterflow observed. In the northern Straits the dominant mode of variability is associated with northward advection of weekly period meanders and frontal eddies, with cross-channel displacements of 5 to 10 km (Duing, 1975; Duing, Mooers and Lee, 1977; Lee and Mayer, 1977; Brooks, 1979; Lee, Schott and Zantopp, 1985; Leaman, Molinari and Vertes, 1987; Johns and Schott, 1987; Schott, Lee and Zantopp, 1988). The FC has also been shown to have a significant seasonal change in current speeds and volume transport in the northern Straits, with maximum currents and transports occurring in summer and minimum in fall (Niiler and Richardson, 1973; Lee and Williams, 1988; Schott et al., 1988).

The Loop Current, on the other hand, displays large northward excursions into the eastern Gulf of Mexico. Northward penetration of the Loop Current tends to be maximum in the late summer and minimum in the fall and early winter, similar to the annual cycle of the FC transport, but with considerable year-to-year variability (Maul, 1977). Cyclonic frontal eddies, similar to those observed north of the Straits of Florida (Lee and Atkinson, 1983), are carried southward by the Loop Current along the seaward edge of the southwest Florida Shelf on 2 to 14 day time scales (Paluszkiewicz et al., 1983; Vukovich and Maul, 1985). These features have diameters of 80 to 120 km and can extend to a depth of about 1000 m. Geostrophic estimates of their swirl speeds range from 100 cm/s in the portion of the eddy nearer the Loop Current to 20 cm/s over the west Florida shelf (Vukovich and Maul, 1985).

Occasionally the Loop Current will shed a large anticyclonic ring or eddy that moves slowly into the western Gulf. The mean period of ring shedding is estimated at 8.5 months (Maul, 1977; Sturges, 1992). Ring diameters can be 300 - 400 km and when shed can cause the flow through the Yucatan Channel to turn directly into the SSF without the presence of a Loop Current (Leipper, 1970; Ichiye et al., 1973; Hurlburt and Thompson, 1980; Muller-Karger et al., 1991). The Loop Current may take several months to reestablish itself after a ring separation. However, at times smaller rings (less than 250 km in diameter) may shed that do not cause a large southward retreat of the Loop Current although the northward penetration distance is decreased (Vukovich, 1995).

Recent Observations

Site characterization of the oceanographic conditions in the Tortugas region is based primarily on results from recent and ongoing investigations of the south Florida coastal waters made by the University of Miami, NOAA/AOML and Science Applications International Corp. of Raliegh, N. C. Long-term moored observations of current and temperature variability have been made along the offshore fringes of the Florida Keys reef tract from Carysfort Reef to the Dry Tortugas by the University of Miami as part of multidisciplinary studies of larval recruitment processes (SEFCAR: South East Florida and Caribbean Recruitment Study) and studies of surface transport processes (SFOSRC: South Florida Oil Spill Research Center) for the period 1989 to 1994. These data are used to make robust estimates of magnitudes and patterns of mean flow and temperature fields in outer shelf coastal waters of the Keys, and are compared to spatial patterns of the mean wind field measured at offshore Coastal Marine Automated Network (CMAN) weather stations, and to the change in coastline orientation of the Florida Keys (Lee and Williams, 1998).

A one year field study of the physical oceanography of the Florida Current was undertaken by SAIC of Raleigh, N. C. with MMS support starting in Nov. 1990. The study involved maintenance of a large moored current meter array, together with seasonal hydrographic cruises and remote thermal imagery obtained from the University of Miami/RSMAS Satellite Facility. The array consisted of moored current meters at 2 to 4 levels from about 100 m to near bottom, on four transects across the FC located at Marquesas Keys to mid-channel, Looe Reef to Cay Sal Bank, Miami to Bimini and Palm Beach to West End (Lee et al., 1995). Additional data on currents, exchange pathways, water mass properties and winds in south Florida coastal waters are obtained from an ongoing study of the circulation and exchange of Florida Bay waters with surrounding regions conducted by University of Miami and NOAA/AOML as part of the South Florida Ecosystem Restoration Prediction and Modeling Program (SFERPM) supported by NOAA/COP.

Seasonal Variation of Water Mass Properties

As part of the SFERPM Program multidisciplinary surveys of water properties of the Keys coastal waters, western Florida Bay and southwest Florida shelf out to the Dry Tortugas are conducted bimonthly from the R/V Calanus (Lee et al., 1998). Nearly simultaneous surveys of similar water properties are conducted monthly through the shallow interior parts of Florida Bay using a shallow draft catamaran. Observations from the first year of these measurements are used to show the strong connectivity of the Tortugas region with the west Florida shelf and Florida Keys; seasonal variability of shelf waters; low-salinity inputs from remote upstream regions; interannual variability caused by enhance wintertime precipitation from the strong El Nino of 1997-1998; and the effects of strong transient wind forcing from Hurricane Georges. Surface salinity distributions from representative seasons are shown in Figs. 3a - 7a. Seasonal salinity sections for the Marco Island to Tortugas transect are presented in Figs. 3b - 7b, and temperature sections for this transect are given in Figs. 3c - 7c.

The typical seasonal pattern of precipitation in south Florida consists of a dry season during winter and spring followed by a wet season during summer and fall (Fig. 7.5). This precipitation pattern can result in hypersaline conditions within the interior lagoons of Florida Bay in the dry season and low-salinity in the wet season (Robblee et al., 1989; Fourqurean et al., 1993). The precipitation pattern of 1998 was unusual in that the strong El Nino of 1997-1998 caused enhance precipitation during the winter/spring "dry season" associated with the passage of moisture laden low-pressure systems from the Gulf of Mexico and decreased precipitation during the summer/fall "wet season" as the El Nino ended and La Nina developed. The salinity patterns on the southwest Florida shelf and Florida Bay show a clear response to the precipitation delivered to the coastal waters through river discharge. A band of low-salinity water was observed along the coast between Cape Sable and Cape Romano in the fall of 1997 that continued to increase in size and fresh water content through the winter/spring wet period (Figs. 3a - 5a), then decreased in size and fresh water content through the summer and fall of 1998 (Figs. 6a and 7a). A similar pattern of enlarging and shrinking fresh water distribution was also apparent within Florida Bay from discharge variations of the Taylor Slough. The general pattern of salinity isopleths is aligned with the local isobaths in a southerly direction, suggesting a net southerly movement and preferred mixing in the direction of flow. The local minimum of salinity near the Shark River mouth indicates a local source of low-salinity from the river discharge. This low-salinity plume, with salinities less than 32 psu, appears to be confined to the region between Cape Sable and Cape Romano and within Florida Bay. A much larger region of decreased salinities ranging from 32 to 36 psu extends from near Cape Romano to the Tortugas. Fresh water discharge through the Everglades is typically less than 10 m³/s and is not sufficient to create this large volume of decreased salinities on the shelf. The orientation of the isopleths toward the north along the west Florida shelf indicates that discharge from larger rivers to the north on the west Florida shelf, or possibly the Mississippi River, are the likely sources.

The salinity of the surface waters of the Tortugas was less than oceanic salinities (36.0 psu) during all surveys and indicates that significant dilution by river discharge into the eastern Gulf of Mexico and transport to the Tortugas region is a common occurrence. Seasonal variability of salinity in the Tortugas did not show a similar pattern as the inshore low-salinity band along the Ten Thousand Islands. Rather, in the Tortugas the surface salinity was highest (35-36 psu) in the winter/spring wet season and lowest in the summer (32-34 psu) dry season, again suggesting low-salinity transport from the north. This pattern of dual fresh water sources from local discharges through the Everglades and low-salinity shelf transport from the north appears to result in low-salinity at the nearshore and offshore portions of the southwest shelf and higher salinity between (Figs. 3a, 4a and 6a). This high salinity region is typically north of the western Keys (Marathon to Marquesas) and south of Cape Romano. Salinities greater than 36 psu can occur in this region (Figs. 3a, 4a), which indicates a significant contribution from local evaporation.

The seasonal variation of vertical structure of water mass properties is shown in the Marco Island to Tortugas CTD sections (Figs. 3b-7b, 3c-7c). The seasonal development of shelf-scale stratification begins in the spring with the decrease in wind mixing and increase in runoff (Figs. 5b,c) and becomes highly stratified in summer with weak wind mixing and transport of low-salinity waters from the north (Figs. 6b,c). At this time the shelf is essentially a two-layered system consisting of a shallow surface mixed layer of about 5-8 m, a sharp thermocline/halocline over about 10 m strata where salinity increased from 34 to 36 psu and temperature decreased from 30 to 20 C, and a well mixed cold, salty bottom layer with salinities greater than 36 psu and temperatures below 20 C. Increased wind mixing in fall begins to break down the stratification (Figs. 3b,c) and by winter the middle portions of the shelf were well mixed vertically (Figs. 4b,c). However, near the nearshore and offshore ends of the section stratification remained throughout the year, although weakest in the winter. Stratification near the Tortugas was caused by subsurface inputs of high salinity (36 psu and greater) waters from the Florida Current and nearsurface inputs of lower salinity shelf waters from the north or entrained along the front of the Loop Current/Florida Current as appears to have occurred in December 1997 (Fig. 4b).

Hurricane Georges passed over the western part of the southwest shelf September 24 and 25, 1998 (Fig. 25). We conducted a multidisciplinary survey in early August (not shown) that was similar to the June survey (Fig. 6a,b,c) in that the shelf was highly stratified with a low-salinity (32-34 psu) surface mixed layer extending from mid shelf to the Tortugas. Approximately 2-3 weeks after the hurricane passage the surface mixed layer was observed to have deepened about 10 m, but significant stratification was still apparent at depths of about 18 to 25 m (Figs. 7b,c). The increased mixing also appears to have increased the near surface salinities by about 2 psu, and homogenized a large portion of the shelf salinity between 34 and 35 psu.

Tortugas Gyre

Water motions and water properties in the Dry Tortugas are strongly influenced by a cold, quasi-stationary, counterclockwise circulation that occurs along the northern boundary of the FC south of the Tortugas. Early observations indicated that this feature was a recirculation of FC water and was associated with an offshore shift or meander of the FC axis (Brooks and Niiler, 1975; Chew, 1974; Vukovich,1988). Cold, cyclonic eddies are observed in satellite derived sea surface temperature fields as common features along the shoreward boundary of: 1) the Loop Current (Paluszkiewicz et al., 1983; Vukovich,1988; Fratantoni et al., 1998); 2) the FC near the Tortugas (Vukovich,1988; Lee et al., 1995; Fratantoni et al., 1998); 3) the lower to mid Keys (Lee et al., 1992); 4) near Miami (Lee, 1975; Lee and Mayer, 1977); and 5) along the western boundary of the Gulf Stream off the southeast U. S. (Legeckis, 1975; Bane et al., 1981; Brooks and Bane, 1981; Lee and Atkinson, 1983; Lee et al., 1991). Eddies of this type are ubiquitous features of frontal zones of all major current systems and hence are known as "frontal eddies" (Lee et al., 1991). Frontal eddies form in the troughs of growing offshore current meanders due to instabilities in the basic flow and density fields. Their growth and decay can take place over days to weeks and their size can range from tens to hundreds of km. Frontal eddies can form anywhere along the current frontal boundary, but there are also regions of preferred growth and decay due to topographic constraints (Lee et al., 1991). Growth regions can occur at the flow exits of channels such as the northern exit of the Straits of Florida and the Yucatan Channel, and also downstream of topographic irregularities like the Charleston bump that penetrates into the Gulf Stream. Decay regions are associated with flows converging onto steep topography as occurs for the Gulf Stream off Cape Hatteras. Typically, frontal eddies travel rapidly downstream with the background current at speeds of 20 to 50 cm/s. However, there are regions where quasi-stationary eddies or "gyres" tend to occur due to the particular geographic setting and flow behavior. At these locations the eddy circulation and water mass structure can endure for long periods of time with little downstream propagation. The Charleston gyre that forms downstream of the Charleston bump when the Gulf Stream is steered offshore and the Tortugas gyre that forms offshore of the Tortugas when the Florida Current is in an offshore position are clear examples.

Lee et al. (1995) has shown that the Tortugas gyre occurs when the southward flowing Loop Current overshoots entry into the SSF and approaches the coast of Cuba before abruptly turning to the east. This offshore position of the Florida Current combined with the strong cyclonic curvature of the flow field can cause a counterclockwise recirculation off the Tortugas, the "Tortugas gyre". A satellite image of the surface thermal patterns in a well-developed gyre is shown in Fig. 8 for March 23, 1995. The strong thermal contrast of the FC front is displaced well offshore near the Tortugas with colder waters located in the interior of the eddy from upwelling of deeper layers. Streamers of warmer waters are entrained around the shoreward side of the feature by the cyclonic circulation and wrap into the eddy center, clearly revealing the circulation pattern. The diameter of this gyre is approximately 110 km and it remained stationary off the Tortugas for about 3 months before subsequently moving eastward into the Straits where it dissipated off the middle Keys in early May (Fratantoni et al., 1998). Recent studies of Tortugas gyres show that their spatial scales range from about 80 to 200 km. They remain stationary off the Tortugas for periods of 12 to 120 days, followed by eastward movement into the Straits at speeds ranging from 4 to 16 km/day, whereupon they decrease rapidly in size and finally dissipate off the middle Keys (Lee et al., 1995; Fratantoni et al., 1998).

Fratantoni (1998) used a combination of numerical model experiments with time series of satellite thermal images to show that the evolution of Tortugas gyres is strongly dependent on the variability of the Loop Current and Loop Current frontal eddies (LCFE). Typically, these gyres remain stationary off the Tortugas until they are forced into the Straits by interaction with an approaching LCFE. The departing eddy is quickly replaced by the LCFE that then becomes stationary off the Tortugas, waiting for impact from the next approaching eddy to send it into the Straits. Fratantoni found two modes of gyre evolution that are shown schematically in Fig. 9. In Mode 1 the Tortugas gyre is forced downstream into the Straits by an approaching LCFE of comparable size as mentioned above. Mode 2 occurs when the southward propagation of a LCFE is interrupted by the formation of a clockwise rotating ring that separates the northern part of the Loop Current and prevents the LCFE from reaching the Tortugas region. In this case the stationary Tortugas gyre must wait for a new LCFE to form on the Loop Current and move into the region of the Tortugas, thereby further increasing the duration of the gyre. There is also a third mode that occurs when a large ring separates and the Loop Current no longer exists, which causes the Yucatan Current to turn directly into the SSF and forces the Tortugas gyre downstream (Lee et al., 1995). This condition may occur about once per year and results in strong eastward flow off the Tortugas that can persist for a time period on the order of one month.

The influence of Tortugas gyre circulation and FC meanders on currents and water mass properties in the Tortugas region are shown in Figs. 10 - 15. The SEFCAR Project conducted several hydrographic surveys in the Tortugas/Marquesas region between May and July 1991 that documented the presence of a cold, cyclonic gyre centered about 40 km south of the Dry Tortugas (Figs. 10a and b). The temperature distribution at 100 m depth shows this cold eddy between the Tortugas and the FC front that extends offshore about 75 km. The dimensions of the gyre are approximately 180 km in alongshore length and 100 km cross-shore, and remained relatively constant over the one month interval between surveys, although the eastern portion of the gyre extended further to the east and across the Marquesas section in the June survey. The cyclonic circulation of the gyre was supported by an upwelling of the thermocline of about 25 m in the center of the feature, near station 21C (Figs. 10 and 11). The surface mixed layer ranged from a minimum of 5 to 15 m near the center of the gyre to about 30 m over the shelf (Fig. 11). The thermocline was located in the upper 150 m, consisting of a thermal strata between 15 and 29 oC. The strongest part of the thermocline occurred in the central part of the gyre between the 23 to 27 oC isotherms. Nutrient profiles show an uplifting of the nutricline to within 50 m of the surface near the gyre center at stations 19C and 21C (Fig. 12a). Nitrate concentrations were about 10 uM at 100 m in the gyre center, and at the edge of the gyre influence the 10 uM level was at a depth of about 120 m (Fig. 12b). Without the gyre induced upwelling nitrate levels would be generally less than 0.1 uM at 50 m and from 1 to 3 uM at the 100 m level (Lee et al., 1992).

Chlorophyll values in early June were greater than 0.5 ug/liter near the gyre center (Fig. 12c). This modest, but significant, enrichment is a probable response to the upwelled nutrients in the region. Abundances of copepod nauplii were also observed to be greatest near the gyre center in the early June survey (not shown). By late June chlorophyll and naupliar values had declined, but ammonia was high in the late June Tortuga section (Fig. 12b), suggesting areas of high grazing and secondary productivity.

Argos satellite tracked nearsurface drifters were deployed on the northern side of the gyre during each hydrographic survey (Figs. 10a and 10b). Each drifter consisted of a PVC cylinder 0.1 m x 2.5 m, which served as the floatation and housed the transmitter, and a canvas drogue (sock with holes) 2 m x 5 m. The 10 m2 area drogue was tethered to extend over depths of 3 to 8 m to track the flow in the surface mixed layer. The drifter design is a slight modification of the well-tested model developed by Bitterman et al. (1979). Trajectories of drifters deployed on May 27 and June 27 are remarkably similar and generally trace the cyclonic circulation of the gyre as indicated by the 100 m temperature pattern (Figs. 10a and 10b). Interestingly, both drifters initially moved toward the WNW across a known mutton snapper spawning area, followed by offshore movement toward the SSE around the gyre, and then shoreward again in the vicinity of Looe Reef. Advective speeds generally ranged from 20 to 40 cms-1 within the gyre and increased to greater than 50 cms-1 after the drifters moved offshore and into the FC front.

A third drifter was deployed on May 30 on the northern side of the gyre, at the same location as the May 27 release (Fig. 13). This drifter spent the first 14 days in a tight recirculation in the gyre interior. On June 12 the drifter broke out of the gyre circulation and entered Rebecca Channel between the Dry Tortugas and Rebecca Shoal where it stayed for about 10 days, undergoing tidal excursions of up to 7 km, but with no net through-flow. The drifter emerged north of the channel on June 22 and proceeded to follow the 20 m isobath east and northward along the southwest Florida shelf to a latitude of 25o 50' N around the end of September. Next, the drifter moved on a general westward course, but with considerable north/south excursions that appear to be the result of cold front passages until about Jan. 5, 1992, whereupon it merged with the Loop Current and was transported rapidly to the south at speeds that sometimes exceeded 100 cms-1. Entering the SSF the drifter appears to have been advected around a well developed cyclonic gyre off the Tortugas and approached within about 40 km to the coast of Cuba. Current measurements from a bottom mounted ADCP off the Tortugas during this period show a strong westward recirculation for Dec. 91 and Jan. 92 (Fig. 14). Altogether this drifter spent approximately 8 months in the southern Straits of Florida and southwest Florida shelf.

Current measurements were made at several sites in the vicinity of the Tortugas during the period from Dec. 1991 to April 1992 using subsurface moorings with standard current meters and a bottom mounted Acoustic Doppler Current Meter (ADCP) to profile the currents over the water column. Location of the mooring sites are shown on Fig. 10. Currents were measured near the southern boundary of the Tortugas Park at 7 m depth in a water depth of 30 m and labeled DT on Fig. 10. Offshore of this site at 200 m depth a bottom mounted ADCP was deployed to profile the currents over the slope (labeled ADCP). Additional moorings were deployed off the Marquesas Keys and Looe Reef (labeled M and L). Selected time series of the 40-hour low pass filtered current vectors are shown in Figs. 14a and b. The gyre observed offshore of the Tortugas during May through July 1991 produced persistent westward flow of 5 to 15 cm/s at the southern Park boundary and westward speeds reaching 40 cm/s closer to the gyre center at the 200 m isobath, with little change in current amplitude with depth. The full year of subtidal currents at the ADCP site off the Tortugas (Fig. 14b) indicates four periods of strong, persistent westward flow from May - Aug. 91, Sep. - Nov. 91, Dec. - Jan. 92 and Feb. - Mar. 92. The duration of these westward flow events ranged from 40 to 108 days, and were interrupted by short burst of intense eastward flow, lasting about 20 to 30 days. The duration of the summer event was 108 days and the fall event lasted 76 days. The fall event was particularly strong with westward speeds reaching 100 cm/s at the ADCP. At the 30 m isobath off the Tortugas the flow was primarily toward the west throughout the measurement period (Fig. 14a).

The eastward propagation of westward flow events as observed in the current meter records at each transect indicates that the Tortugas recirculation features spread eastward and caused cold flow reversals off Marquesas from about May 24 to Sept 5 and again from Sept. 13 to Oct 29 (Fig. 14). These features moved downstream at approximately 5 km/day. The cold recirculations caused westward flow near Looe Reef from about Aug. 28 to Sept. 17 and Oct. 27 to about Nov. 6. Also the occurrence of the gyre offshore of Looe Reef was coupled to flow increases (decreases) on the eastern (western) side of Santaren Channel and a temperature drop throughout the channel, indicating that the offshore displacement of the FC axis caused a splitting of the flow around Cay Sal Bank.

The durations of the Tortugas recirculations as observed at the Marquesas section were about 105 and 45 days for the summer and fall events, respectively. At the Looe Reef transect these gyre durations decreased to 21 and 11 days, respectively. The shorter gyre durations at the eastward locations indicates that the gyre decreased in size as it approached the northward turn of the Straits of Florida, where the channel width decreases by about a factor of two, due to the presence of the Cay Sal Bank. Indeed, satellite thermal imagery show that for the April event, gyre dimensions decreased from about 90 x 150 km (cross-shore x alongshore distance) off the Tortugas to 70 x 120 km off the Marquesas and approximately 60 x 100 km off Looe Reef. Fratantoni et al (1998) has shown that Tortugas eddies in the Florida Keys shrink to approximately 55% of their initial size off the Tortugas and increase their downstream translation speed from about 5 km/day off the Tortugas to 17 km/day in the Middle Keys.

In the SSF the dominant period of FC meanders occurs in a 30 to 70 day band and the FC front can be displaced up to 100 km offshore during a meander. The cross-channel position of the axis appears to be strongly influenced by the manner in which the FC enters the southern straits from the eastern Gulf of Mexico, as well as by the eddies that travel downstream along the Loop Current cyclonic front. Satellite derived surface temperature patterns show that in December 1990 the Loop Current was nonexistent and the warm flow through the Yucatan Channel turned abruptly to the east and entered the SSF with little penetration into the eastern Gulf of Mexico (Fig. 15a). This anticyclonic turning of the flow caused the northern boundary of the FC to extend close to the Tortugas and lower Florida Keys, resulting in strong downstream flows at the inshore moorings on the Marquesas transect. As the Loop Current grew northward into the eastern Gulf of Mexico the southward limb of the flow along the seaward edge of the southwest Florida shelf appeared to overshoot the entrance to the SSF before turning abruptly to the east off the coast of Cuba. The southward overshooting of the Loop Current, combined with the strong cyclonic curvature vorticity from the sharp eastward turning, appeared to spin-up a cold cyclonic gyre offshore of the Dry Tortugas. Further enhancement of the gyre appeared to occur with encounters of Loop Current frontal eddies that propagate into the region (Fratantoni et al., 1998).

Mean Wind, Current and Temperature Patterns

Five years of moored observations of coastal currents and temperature along the outer shelf of the Florida Keys from Carysfort Reef to the Dry Tortugas over the period April 1989 to April 1994 are combined with local CMAN wind records and current observations from offshore waters to describe mean distributions and seasonal cycles of the current and temperature fields. These data were collected as part of the NOAA supported study of South East Florida and Caribbean Recruitment (SEFCAR), and U. S. Coast Guard supported studies of surface transport processes by the South Florida Oil Spill Research Center (SFOSRC). Mean distributions and seasonal cycles of currents and temperature along the outer shelf of the Keys are shown to result from the combined influences of Florida Current forcing in the form of transient gyres, eddies and meanders, together with seasonal variations of Florida current transports, prevailing easterly winds and the cyclonic curvature of the coastline. Previous studies of current variability in the Florida Keys coastal waters indicate that in the cross-shelf direction there are two distinct flow regimes, associated with responses to different forcing mechanisms. The first flow regime consists of the nearshore and Hawk Channel regions, where flow variability is controlled by tidal and local wind forcing (Lee, 1986; Pitts, 1994). Tidal currents are primarily in the alongshore direction (except for areas immediately adjacent to tidal passages between the Keys) and can account for 20 to 50% of the total current variance, but exhibit no significant tidal residual flow (Lee, 1986; Pitts, 1994). Low-frequency variability, primarily from local alongshore wind forcing, accounts for the remaining 50 to 80% of the total variance. The wind response appears typical of other shallow coastal zones with long, straight coastlines (Lee and Mayer, 1977; Csanady, 1982) and consists of significant alongshore flow events directly forced by local alongshore winds, which are primarily responsible for seasonal mean flows (Lee, 1986; Pitts, 1994). However, the Florida Keys coastline curves from an east-west to north-south orientation and this can produce an interesting pattern of diverging alongshore flows forced by persistent easterly (westward) winds, with westward currents prevailing in the lower Keys and northward currents more common in the upper Keys (Pitts, 1994). Seasonal shifts in wind directions from southeasterly in summer to northeasterly in fall and east-northeasterly in winter and spring can cause seasonal shifts in mean alongshore current directions in the upper Keys that follow the seasonal wind shifts (Lee, 1986), whereas westward flow persists throughout the year in the lower Keys (Pitts, 1994).

The second flow regime is the outer shelf region where flow varibility is more strongly influenced by the Florida Current and wind forcing, and only to a lesser degree by tidal forcing. In the northern Keys and in the continuation of the coastal zone from Miami to Palm Beach, flow in the outer shelf is directly influenced by the passage of Florida Current meanders and eddies on weekly time scales. Low-frequency fluctuations account for approximately 75% of the total current variance and over 90% of the total temperature variance (Lee, 1975; Lee and Mayer, 1977; Lee and Mooers, 1977; Zantopp, Leaman and Lee, 1987; Lee, 1986; Lee et al., 1992). Local wind forcing accounts for only about 10% of the total alongshore current variability (Lee and Mayer, 1977). Close proximity to the Florida Current front at these locations also results in strong, persistent northward mean flows with significant vertical shear. In the southern Keys, the Florida Current influence occurs on a 30 to 60 day time scale due to the passage of cyclonic gyres that develop off the Dry Tortugas, in addition to the shorter weekly time scale of variability due to the occurrence of spin-off eddies frequently found in the northern Keys and off the coast of southeast Florida (Lee et al., 1992; Lee et al., 1994). Also, local alongshore wind forcing becomes more significant in the lower Keys due to the east-west orientation of the coastline, which aligns with the prevailing wind direction. The current response to the persistent westward alongshore winds results in downwelling along the coast, with onshore (offshore) flows in the upper (lower) Ekman layers and westward barotropic alongshore flow (Lee et al., 1992).

Subsurface, taut-wire current meter moorings have been maintained along the 30 m isobath between the Dry Tortugas and Key Largo for durations of one year or longer during the 5 year period encompassing April 1989 to December 1994. Currents and temperature were measured with General Oceanics current meters at depths of 7, 17 and 27 m. Mark I and II current meters were used with burst sampling over a 32 second period to reduce surface wave noise. The Looe Reef site has been maintained continuously and there are now over 9 years of data at this site. The upper and middle Keys sites at Carysfort and Tennessee Reefs were in operation from April 1989 to April 1991. The western Keys site off the Marquesas was active for more than 3.5 years, from April 1991 to November 1994, and the Dry Tortugas site was maintained for one year, from April 1991 to May 1992. In addition bottom-moored ADCP's were deployed offshore of the Dry Tortugas at the 200 m isobath from April 1991 to May 1992, and at the 150 m isobath off Looe Reef for the periods August to November 1989; April 1990 to April 1991; and July 1993 to November 1994, to measure variations of the Florida Current. Wind records were obtained for the same 5 year period from the offshore CMAN stations positioned on lighthouses at Fowey Rocks, Molasses Reef, Sombrero Reef, Sands Key and the Dry Tortugas.

Averages over the total record lengths, as well as annual averages of current, temperature and wind time series, are given in Tables 1 and 2. Annual averages are computed as the averages of monthly means. Annual averages are more representative of the long term mean conditions in the Keys coastal waters, where significant annual cycles could cause discrepancies in short record length means that end at different phases of the annual cycle. Also reported are the instrument locations, rotation angles used to convert velocity vectors into the isobath coordinate system, the total record lengths, standard deviations (S.D.), standard error of the means (S.E.), and the percent of the total variance due to low-frequency (40 hlp filtered) and high frequencey (3 - 40 hour bandpass) variations. The standard error is computed from: S.E. = S.D./(d.f.)^1/2, where d.f. is the degrees of freedom, given by the record length (days) divided by the time scale (Press et al., 1992). The standard error of the mean currents and temperatures were computed using a 10 day time scale for the upper Keys sites (typical time scale of Florida Current meanders and frontal eddies in the upper Keys) and a 60 day time scale for the lower Keys sites (typical time scale of Florida Current meanders and gyres in the lower and western Keys, from Lee et al., 1994). A ten day time scale typical of the most energetic period for wind forcing was used to compute S.E. of the wind records. Annual statistics are not shown when there were fewer than 8 calendar months of data.

Generally the record length averages agree quite closely with the annual means, with differences usually less than 1 cm/s for the currents, 0.3^oC for temperature and 0.2 m/s for winds, and no significant differences in sign. This indicates that the record lengths are sufficiently long to produce reasonably stable mean conditions. Robust estimates of mean temperature were made for all sites with standard errors less than 0.7 ^oC. Mean alongshore flows were well resolved at those locations with strong mean currents, e.g., the upper Keys and the offshore ADCP site off Looe Key, where standard errors were typically less than half of the means. Whereas, at coastal sites with weak mean flows the standard errors ranged from 2 to 5 cm/s and could be several times the mean. Alongshore coastal currents at the Dry Tortugas site show a significant westward mean flow of about -8.0 cm/s in the upper layer with little subtidal variability (Table 1, standard deviations of low-frequency alongshore current records are about equal to the mean). The most uncertain mean flow occurred at the offshore ADCP site at the Dry Tortugas, where year-long mean flows were about 1.0 cm/s and standard errors were 12 to 14 cm/s. Standard deviations of the alongshore currents at this site were about 40 cm/s, and reflect the location of the site, which at times is characterized by strong westward flows during Tortugas gyre events (which can last 2 to 3 months), and at other times by strong eastward flow events, which can last for about one month when the Florida Current front meanders onshore (Lee et al., 1994). Standard errors of cross-shelf currents are typically about the same magnitude as the means, except again for the Dry Tortugas ADCP site where the standard errors are about twice the mean. Annual mean winds are well resolved over the five year period and are nearly equal in magnitude, even though record lengths varied at the different sites from 2 to 5 years (Table 2). Mean winds were toward the west at about 2 to 3 m/s with standard errors generally less than 0.3 m/s.

Annual mean current and wind vectors from each site are shown in Figure 16 for the five year period 1989 - 1994. Mean wind vectors are toward the west at all sites and are of similar magnitude. The observed small deviations in mean wind directions are more likely due to differences in record lengths rather than statistically meaningful direction shifts. Subtidal wind records normally show a high degree of spatial coherence over south Florida and the Keys due to both the zonal orientation of the CMAN stations in a westward wind regime, and to the large spatial coherence scales of synoptic wind forcing events. The cyclonic curvature of the Keys coastal boundary results in mean wind vectors being oriented mostly cross-shore in the upper Keys, and alongshore from the middle and lower Keys out to the Dry Tortugas.

Mean coastal currents show the influence of both the persistent westward winds, the Florida Current and coastal gyres. In the upper Keys strong mean downstream flows occurred with considerable vertical shear, indicating direct influence from the nearby Florida Current front. At the Tennessee Reef middle Keys mooring site mean currents show considerable decrease in amplitude and change in direction with depth, i.e. upper currents were strong in the downstream and onshore directions and near-bottom currents were predominantly offshore. In the lower and western Keys the mean flow was westward and steadily increased from Looe Reef to the Dry Tortugas. A westward coastal countercurrent is to be expected from the westward wind forcing. However, the westward intensification of the countercurrent appears to be related to the prolonged duration of the Tortugas gyre off the Dry Tortugas. Lee et al. (1994) show that this cyclonic gyre is a persistent feature off the Tortugas causing westward flow on the northern side of the gyre that can last up to three months. The presence of the gyre causes the Florida Current front to be displaced further offshore on the mean than at any other location in the Straits. The combined effect of the gyre circulation and the offshore displacement of the Florida Current results in the weak eastward mean flows observed at the ADCP site off the Dry Tortugas (Fig. 16). The more shoreward mean position of the Florida Current front off Looe Reef results in the strong downstream mean flows observed at the nearby ADCP site on the Pourtales Terrace.

The alongshore distributions of alongshore and cross-shore mean flows and temperature at the 30 m isobath are shown in Figure 17. There is a mean alongshore flow divergence, with strong downstream (northward) flow in the upper Keys and upstream (westward) flow in the lower Keys out to the Dry Tortugas. Compensation for this flow divergence may occur as inflow from the Florida Current, as suggested by the strong onshore mean flow at Tennessee Reef (Fig. 17b). Additional compensation may occur as inflow from the Gulf of Mexico and Florida Bay through the Keys passages, as shown by Smith (1994) and Lee et al. (1998). In the upper Keys at Carysfort Reef a vertically sheared, strong northward mean flow occurred with an onshore component over the entire water column. Downstream mean flow was nearly 20 cm/s in the upper 20 m and decreased sharply to about 10 cm/s near the bottom. Onshore mean flows were about -1 to -2 cm/s (negative is onshore) at all depths. A strong northward mean baroclinic flow at this location indicates a close proximity to the cyclonically- and vertically-sheared Florida Current front, where downstream flows increase rapidly with distance offshore and above the bottom. The observed onshore mean flows over the total water column suggest a shoreward convergence of the Florida Current front, which occurs following the cyclonic northward turn of the current in the middle Keys. Onshore mean flows were maximum in the upper layer at the Tennessee Reef site, which could be explained by the shoreward movement of the Florida Current in this area in combination with onshore Ekman transports and partial compensation for the mean alongshore flow divergence.

The trend of increasing upstream mean flows from the lower Keys to the Dry Tortugas occurs due to the combined effects of alongshore wind forcing and the increased persistence of coastal gyres offshore of the Dry Tortugas. The east-west orientation of the coastline and bottom topography in the lower Keys is more aligned with the prevailing westward winds resulting in significant alongshore wind forcing compared to the upper Keys where cross-shelf winds prevail (Fig. 16). Westward wind forcing will result in nearly barotropic westward alongshore flow, typical of the Looe Reef area, combined with onshore flow in the upper Ekman layer and offshore flow in the lower layer, which occurred on the average from Looe Reef to the Marquesas (Figs. 17a and 17b). The longer durations of cyclonic coastal gyres off the Dry Tortugas (60 to 100 days) enhances the magnitude of the upstream mean flow, compared to the Looe Reef area where gyre durations are about one month or less (Lee et al., 1994). The gyre circulation off the Dry Tortugas also appears to result in a mean offshore flow in the upper layer at this site, overcoming the onshore Ekman transports typical of this downwelling coastal regime. Without the gyre influence the mean alongshore and cross-shore flows would be nearly uniform along the east-west oriented coastal zone of the lower Keys. The persistence of cyclonic gyres off the Dry Tortugas and their accompanying upwelling also appears to have a significant influence on the mean temperature field at the 30 m isobath, resulting in increased stratification and colder lower layer mean temperatures in the west (Fig. 17c). There is also a trend of increasing upper layer temperatures from the upper Keys to western lower Keys.

Tidal and Inertial Variability

The percent of total variance of coastal current components and temperature due to high-frequency (tidal and inertial) motions are shown in Table 1 for all stations at the 30 m isobath in the % variance column labled 3-40 hlp. This represents the percent of the total variance due to fluctuations in the period band of 3 to 40 hours, ie the tidal to inertial motion period band. The 40 hlp column represents the percent of the total variance due to subtidal fluctuations. Taken together the two columns represent 100% of the variance. Table 1 indicates that tidal motions account for about 30% of the total alongshore current variance at upper and mid-water positions at all Keys coastal sites except for the Dry Tortugas where about 70% of the variance is explained by tidal and inertial motions. In the lower layer approximately 50 to 40% of alongshore current variance is accounted for by tidal motions, again except for the Dry Tortugas where 90% is due to tidal and inertial fluctuations. In the region from Key Largo to Key West approximately 50 to 70% of the total variance of cross-shelf currents is explained by tidal and inertial motions. In contrast, at the Marquesas and Dry Tortugas sites, which are more openly connected to the Gulf of Mexico, tidal and inertial motions account for about 85% of the total cross-shelf current variance. Temperature variability is dominated by low-frequency fluctuations at all sites, accounting for greater than 95% of the total variance leaving only about 5% related to tidal and inertial variations.

In the western Keys, which are openly connected to the Gulf of Mexico, coastal current variability is strongly influenced by tidal and inertial motions due to significant tidal interaction between the Gulf and Atlantic at diurnal and semi-diurnal periods (Zetler and Hansen, 1970). The semi-diurnal variability results from a semi-diurnal tidal wave propagating southward from the Atlantic through the Straits of Florida. The diurnal variability is due to the diurnal tidal wave of the Gulf of Mexico that oscillates near the natural period of the basin. The inertial period at this latitude is also close to the diurnal period, which can provide further high-frequency current variance.

Tidal currents were measured near the center of Rebecca Channel in 1991 as part of a Minerals Management Services study of the physical oceanography of the Straits of Florida (SAIC, 1992). Tidal currents were aligned with the Channel axis and were primarily semi-diurnal with amplitude of the M2 tidal constituent of about +/- 20 cm/s. The diurnal components (K1 and O1) had amplitudes of about +/- 7 cm/s each. North of the Tortugas on the southwest Florida shelf, recent current measurements (mooring locations are shown in Figure 25) reveal that tidal currents are primarily semi-dirunal and oriented east-west with amplitudes reaching +/- 40 cm/s. Strong tidal currents in the Tortugas region can result in significant mixing and dispersion of planktonic particles, but little net advection due to the oscillatory nature of tidal currents. For significant spatial displacement and net advection of particles to occur requires transport by background currents with longer periods, ie subtidal currents generated by wind and Florida Current mechanisms.

Annual Cycles

Five years of current meter data from the Looe Reef site and CMAN wind records for the period April 1989 to April 1994 were used to compute monthly averaged time series to investigate seasonal variability. Ensembles of monthly average alongshore and cross-shore wind stress components are shown in Fig. 18 for Fowey, Molasses, Sombrero and Sand Key stations. Records from the Fowey Rock Light station are characteristic of an upper Keys site with no isobath rotation, and the Sombrero and Sand Key stations are representative of lower Keys sites with isobath rotations of 73 and 90 degrees, respectively. Standard errors of the monthly means are small in comparison to the scale of the seasonal cycles, giving robust estimates of seasonal variability with the available data. The monthly average alongshore wind stress was consistently negative (toward the west or southwest) at the middle and lower Keys sites, and shows similar annual cycles of maximum westward and southwestward wind during the fall to early winter period of Oct. to Jan. and minimum alongshore wind during the summer (June to Aug.). There is also a weaker, secondary maximum in the spring (April and May). The northern Keys site (Fowey) showed a similar pattern with maximum southward wind stress in the fall and maximum northward wind in the summer when southeasterly winds prevail. Alongshore wind stress was consistently stronger in the lower Keys than in the upper Keys due to the curvature of the coastal zone. However, this difference becomes smaller in the fall (Oct. to Nov.) when persistent winds from the northeast occur. The effect of isobath curvature is clearly shown by comparison of the seasonal patterns of cross-shore wind stress (Fig. 18b). At the upper Keys site cross-shore wind stress was onshore for all months with little seasonal variation and was consistently stronger than at the lower Keys site. Whereas, at the lower Keys sites cross-shore winds show a significant seasonal pattern of onshore winds prevailing from April to Sept., with a maximum in the summer, and then shifting to offshore in the fall when winds from the northeast prevail.

Monthly average alongshore and cross-shore currents and temperature from the Looe Reef site are shown in Figure 19. Standard errors of the monthly means are only shown for the instrument at 17 m to make the figure less cluttered. However the 7 and 27 m means have similar error bars and indicate that the seasonal cycles shown are statistically well resolved. Monthly mean alongshore currents have a significant seasonal cycle with maximum eastward flow occurring in late winter/early spring (Feb. to Mar.) and again in late summer (July to Sept.). Maximum westward flow occurred in the fall (Oct. to Nov.). This pattern was similar at all depth levels, except the amplitude was weaker in the lower layer and the spring eastward maximum was not observed there. The annual cycle of alongshore currents resembles the annual cycle of alongshore winds (Fig. 18), especially in the lower layer where westward flow persists. The fall maximum of westward flow occurred during the time of maximum westward wind stress. The summer maximum of eastward flow occurred during the period of minimum westward wind stress, which suggests that the seasonal alongshore currents are also influenced by the seasonal cycle of the Florida Current with maximum downstream transport in the summer and minimum in the fall (Niiler and Richardson, 1973; Lee and Williams, 1988). Also the secondary maximum in eastward flow during late winter/early spring is at odds with the seasonal alongshore winds and suggests a Florida Current source.

The seasonal cycle of cross-shore currents appears to be the result of Ekman response to alongshore wind forcing (Fig. 19b). Cross-shore flows in the upper and lower layers are out of phase and closely follow the seasonal cycle of alongshore winds. Both maximum onshore (offshore) flows in the upper (lower) layers occurred at the time of maximum westward winds in the fall and spring (Fig. 18a). Evidence that the cross-shore flows were directly wind forced is also shown by the weak cross-shore flows in the summer when alongshore winds were weak and alongshore currents were strong eastward. At mid-depth monthly averaged cross-shore flow is near zero.

The annual cycle of monthly averaged water temperatures at the Looe Reef site has a maximum in summer of about 30^oC and minimum in winter near 24^oC that clearly follows the annual cycle of air temperature in the Keys (Fig. 19c). The seasonal change in temperature stratification also appears to be atmospherically forced with minimum temperature difference between the upper and lower layers occurring in the fall and early winter during maximum wind mixing and maximum temperature difference occurring in late winter/spring and summer during weaker wind mixing periods.

Seasonal wind patterns in the Keys result in weak southeasterly winds in summer, stronger northeasterlies in the fall and moderate to strong east to northeasterlies in the winter and spring (Fig. 18). Because of the curving coastline, the fall season with increased winds towards the west and southwest, becomes the dominant alongshore wind forcing season for the entire Keys from the Dry Tortugas to Key Largo (Fig. 18). Seasonal variations of currents off Looe Reef determined for the same five year period as the seasonal winds are significantly correlated with alongshore winds. Maximum westward monthly mean alongshore flows, together with the strongest downwelling-type Ekman cross-shelf mean flows, occurred simultaneously with the fall annual westward peak in alongshore wind forcing. A similar seasonal pattern is found in the monthly averaged and seasonal averaged alongshore currents for the entire length of the Florida Keys (Fig. 20). Seasonal averages of the alongshore currents provide some smoothing to small irregularities in the monthly means caused by differences in deployment periods and record lengths and clearly show the seasonal pattern of maximum downstream flow in the spring and summer and maximum upstream flow in the fall at all sites that matches the seasonal pattern of alongshore wind stress. This pattern is displaced in a downstream direction at Carysfort and Tennessee sites due to the strong influence of the Florida Current in the upper Keys. The Looe Reef site represents a transition region between the Florida Current dominated upper Keys and wind and gyre dominated western Keys and as a result the seasonal current pattern is downstream in spring and summer and upstream in fall and winter. The pattern shifts to a upstream orientation at the Marquesas western Key site where winds are primarily toward the west and gyre influence becomes stronger.

Regional and Remote Interconnections

The Tortugas region is strongly connected to coastal ecosystems of the Florida Keys, western Florida Bay and Ten Thousand Islands, as well as more remote regions of the shelf environments of the eastern Gulf of Mexico, Yucatan and the western Caribbean. The trajectories of satellite tracked surface drifters deployed in the Shark River discharge plume clearly show strong linkages between the Tortugas and Florida Bay and Florida Keys (Figs. 21 a and b). There are two typical exchange pathways coupling these regions. The most persistent route is southeastward through western Florida Bay and the Keys tidal passages then westward along the reef tract to the Tortugas (Fig. 21a). The route through western Florida Bay is driven primarily by local wind forcing and the sea level slope between the Gulf of Mexico and the Atlantic (Lee et al., 1998; Wang, 1998). The westward trajectory in Hawk Channel and the reef tract is sustained by local alongshore westward wind forcing that is enhanced toward the west by recirculating gyres and eddies north of the FC (Lee and Williams, 1998). This coastal countercurrent system is further indicated by the warm filaments from the FC front entrained toward the west along the outer reefs at the time of the westward movement of the drifter through this region (Fig. 22). It takes one to two months for drifters released in the Shark River plume to reach the Florida Keys, and then less than two weeks to reach the Tortugas region due to the increase flow in the coastal countercurrent. However if winds have a significant southerly component then drifters entering the Keys coastal waters through the tidal channels in the middle Keys will turn toward the north and may become entrained in the strong northward FC. A more direct route between Everglades discharge and the Tortugas occurs during east and northeast wind forcing when near surface flows on the southwest Florida shelf are toward the southwest (Fig. 21b). The transport time scale to reach the Tortugas is approximately one month. This southwest route is more typical of the fall season when east and northeast winds prevail, but can occur during any season if east and northeast winds persist. A case in point is the anomalous northeast winds during the El Nino winter of 1998 that caused a surface drifter released in the Shark River plume on Feb. 5, 1998 to reach the Tortugas via the southwestern route in approximately two months (Fig. 23a), whereupon it was entrained into a Tortugas gyre and recirculated south of the Tortugas (Fig. 23b).

The Tortugas region is also highly connected to west Florida shelf waters by the combined influence of the counterclockwise circulation in the Tortugas gyre, wind forced shelf circulation and Loop Current influences. Surface drifter patterns reveal northward movement of gyre return flows onto the west Florida shelf in the vicinity of the Tortugas to be a common occurrence (Figs. 13, 23a and 24). The northward flow is primarily a response to the prevailing southeasterly winds and has been observed to extend along the inner shelf to at least 26o N before turning offshore (Fig. 13). The shelf offshore movement is again a result of wind forcing that shifts seasonally to northeast winds in the fall. Offshore movement continues until the drifter is entrained in the southward flow of the Loop Current and returns to the Tortugas region, whereupon it can again be entrained by the gyre circulation and repeat once again the exchange route with the west Florida shelf. Drifters have been observed to repeat this circuit several times over 5 to 8 month periods before ejection into the FC (Fig. 24).

The primary subtidal current response to wind forcing on wide, shallow continental shelves, such as the west Florida shelf, is a strong alongshore current directly forced by coherent, synoptic scale alongshore winds (Csanady, 1978; Beardsley and Butman, 1974; Lee et al., 1986). During the summer prevailing winds are weak from the southeast, which can drive a weak northward flow along the west Florida shelf. However, during the winter and spring cold fronts pass over the shelf at periods of 5 to 10 days and cause strong wind forcing from the northwest and north that can force southward shelf currents toward the Tortugas.

As part of a recent University of Miami and NOAA/AOML study of the circulation on the southwest Florida shelf and Florida Bay conducted under the auspices of the South Florida Ecosystem Restoration, Prediction and Modeling Program (SFERPM) of NOAA/COP an array of Acoustic Doppler Current Profilers (ADCP’s) are being maintained on the southwest Florida shelf (Lee et al., 1998). Locations of these moorings are shown in Fig. 25. The ADCP array was located within the inner part of the southwest Florida shelf at the 13 and 6 m isobaths. Subtidal vector time series from the first six months deployment of these instruments over fall, winter and spring seasons are shown in Fig. 26, together with wind vectors from the Sombrero Light CMAN station (Fig. 16). All records were filtered with a 40 hour low pass filter to remove the tidal fluctuations. Tidal current amplitudes were 20 to 40 cm/s and dominate the total current variance, but do not contribute significantly to net water movements and are therefore removed from the records to clearly display the subtidal motions. Subtidal current fluctuations are primarily in the alongshore direction as a direct response to alongshore wind forcing events. Coherence between wind and current fluctuations in the period band of 2 days to 2 weeks is very high at the 95% significant level, with the currents lagging the Sombrero winds by about 12 hours. Presumably this lag time would be shorter if wind records were available closer to the current measurement sites. However, Lee and Williams (1998) have shown that wind variability over the Keys as measured at CMAN sites (Fig. 16) is highly coherent and useful for comparison to the currents on the southwest shelf. Amplitudes of the wind driven currents ranged from about 6 to 18 cm/s and were only slightly stronger near the surface with little change of direction with depth, indicating a barotropic or uniform response to wind forcing, as found previously for the west Florida shelf further north of our study area (Mitchum and Sturges, 1982).

Interestingly, the amplitude of southward flow events appears to be significantly larger than the northward flow events, indicating the presence of a significant southward background current or mean flow. This mean southward offset does not appear to be wind induced for the amplitudes of southward and northward wind events are approximately equal. The six month mean flows from all sites are shown on Fig. 16. A southward mean flow of about 3 cm/s occurred nearsurface at the 13 m stations and decreased to about 1.5 cm/s near the bottom. The physical mechanism controlling the strength and direction of this mean flow is not understood at present. However, it is quite possible that the southward flowing Loop Current along the outer part of the west Florida shelf can force a southward flow in the middle and inner shelf regions as well. The physical mechanism involved is either the horizontal frictional drag of the outer shelf current or the generation of an alongshore pressure gradient over the shelf by the southward sloping sea level of the Loop Current. These mechanisms are being investigated.

Connection of the Tortugas region with remote areas of the eastern Gulf of Mexico was recently verified by discovery of large volumes of Mississippi River (MR) water in the Florida Keys (Ortner et al., 1995; Gilbert et al., 1996). Multidisciplinary surveys during the period 10-13 September 1993 revealed a band of anomalously low-salinity water embedded in the FC and adjacent coastal waters extending from Key West to Miami, a distance of 260 km (Fig. 27). Surface salinity values as low as 31 psu were found, where typically surface salinity is greater than 36 psu. The average offshore distance of the band was approximately 40 km. A CTD section off Looe Reef shows the band extended to approximately 20 m in depth (Fig. 28). The estimated volume of the band is approximately 33.3 x 10¹⁰ m³ for the Key West to Miami region, thereby requiring about 1.2 x 10¹⁰ m³ of fresh water to mix with oceanic waters to produce the low-salinity band. Biological and chemical indicators within the band, together with its large volume, suggest the 1993 MR flood as the only conceivable source (Ortner et al., 1995). The mechanism for transport of MR water into the Straits of Florida is indicated by Figure 29, a composite of the sea surface temperature fields derived from satellite images of August 5-7, 1993. MR discharge waters are indicated by the red band extending from the MR delta. The MR discharge appears to be entrained along the eastern front of the LC (green) and transported in the Straits of Florida. A satellite tracked surface drifter, initially deployed on the Texas shelf as part of a Minerals Management Services study, also appears to be entrained in the MR plume on August 19 and arrived in the middle Florida Keys September 10, further substantiating the strong connection between regions by the Loop Current pathway. Later, this drifter reached Cape Lookout, N.C. on September 22, when simultaneous shipboard measurements revealed anomalously low-salinity water along the Gulf Stream front and outer shelf (Tester and Atkinson, 1994). Salinity time series from CMAN stations in the Keys indicate that the low-salinity water remained in the Keys coastal waters for approximately 3 months (Gilbert et al., 1996). The arrival of the band occurred first in the upper Keys, where the FC front was located close to shore in mid August, followed by delayed occurrence in the middle Keys and Tortugas regions as the FC front meandered on shore in those areas (Fig. 30). The transport of MR water to the Florida Keys appears to be the result of several independent processes occurring simultaneously. The unusually high discharge of MR water occurred at a time when the Loop Current was in an extreme northerly position. The combination of summer heating and the river discharge vertically stratified the shelf waters near the delta inhibiting mixing of the MR plume with ambient shelf waters. Persistent upwelling favorable winds transported the MR plume offshore to the southeast where it was entrained by the Loop Current, and transported to the Straits of Florida along its shoreward front.

Implications for Larval Transport - The Recruitment Conveyor

The combination of the above mentioned physical processes tends to form a recirculating retention zone and recruitment pathway for pelagic larvae spawned in the Florida Keys coastal waters or foreign larvae transported from remote sources. This recruitment conveyor system is shown schematically in Figure 31. The four primary physical processes that drive the system are:

1. The Florida Current/Loop Current - The Florida Current/Loop Current forms the offshore leg of the conveyor. Rapid downstream transport occurs in currents that can reach 100 to 200 cm/s. Larvae can be transported great distances from remote upstream sources in the eastern Gulf of Mexico and Caribbean Sea. This can be particularly significant for species with long pelagic larval stages, such as spiny lobster larvae that can remain in the plankton for 6 to 12 months (Lewis, 1951; Sims and Ingle, 1966). The shoreward front of the Florida Current/Loop Current is an area of nearsurface current convergence. Therefore, both larvae and their planktonic food source will tend to be concentrated together in the frontal zone (Yeung, 1996). Onshore meanders of the front can transport larvae closer to the coastal zone and settlement habitat. Also the mean shoreward displacement of the front in the middle and upper Keys will carry larvae closer to settlement habitat in this region. Larval detrainment from the front to the coastal zone can occur through small-scale cross-frontal mixing, eddy circulations, surface Ekman transports and swimming in late stage larvae. Increased abundances of lobster and conch larvae have been observed near the outer reefs when the Florida Current front was located in a nearshore position (Yeung, 1996; Stoner, Mehta and Lee, 1997). Eddy circulations in small-scale frontal eddies and Tortugas gyres have been shown to aid exchange of Florida Current and coastal species of fish larvae and pink shrimp larvae (Limouzy-Paris et al., 1996; Criales and Lee, 1995).
2. The Tortugas gyre - The cyclonic circulation of the Tortugas gyre and its evolution into smaller gyres in the lower Keys provides a mechanism to entrain newly spawned larvae that can be retained in the gyre circulation for up to several months or escape the gyre on one of its shoreward circuits (Lee et al., 1994). Pink shrimp larvae have been shown to take advantage of this pathway (Criales and Lee, 1995). The Tortugas gyre also enhances food availability through increased primary production from upwelling of about 2 m/day and concentration of microzooplankton (Lee et al., 1992; 1994). Early stage spiny lobster larvae have been shown to be concentrated within the gyre and late stages near the Florida Current front and in the gyre interior (Fig. 32), indicating that the gyre circulation functions to retain both locally spawned lobster larvae and foreign recruits.
3. Shoreward Ekman transports - Onshore surface Ekman transports prevail throughout the region from the Dry Tortugas to the middle Keys due to persistent westward winds and east-west orientation of the coastal zone in this region. Spiny lobster larvae tend to be distributed in the upper mixed layer above the thermocline (Yeung and McGowan, 1991; Yeung et al., 1993), where onshore Ekman transports in the upper layer can result in concentration in the Florida Current front and detrainment into the coastal zone. Shoreward Ekman transports are further enhanced near the Florida Current axis, where surface currents and winds are opposed, significantly increasing the surface stress and onshore Ekman transport (Rooth and Xie, 1992) and resulting in concentration of larvae along the front or in the interior of a coastal gyre.
4. Coastal countercurrent - The westward flowing coastal countercurrent described above (Figs. 16 and 17a) as resulting from the combined influences of downwelling winds and coastal gyres provides the primary return leg of the recruitment conveyor. This feature can extend from the middle Keys to the Dry Tortugas. Its northern extent is limited by the curving coastline that cause the prevailing westward winds to change from an alongshore orientation in the lower Keys to onshore in the upper Keys. The maximum northward penetration of the countercurrent occurs in the fall when southwestward winds prevail. These fall winds are oriented alongshore in the middle Keys, and have a southward component in the upper Keys, which could result in a countercurrent extending the entire length of the Keys from the Dry Tortugas to Key Largo. As shown here (Figs. 16 and 17a), the coastal countercurrent is prevalent in the outer shelf, seaward of the reef tract, but it is also believed to extend shoreward to Hawk Channel and the nearshore waters (Lee, 1986; Pitts, 1994). Cross-shore flows in the countercurrent are onshore in the upper layer and offshore in the lower layer in the lower Keys due to an Ekman response to the westward (downwelling) winds (Fig. 17b). However the cross-shelf flow can shift to offshore in the Dry Tortugas region due to influence from the cyclonic circulation in the gyre. This gyre circulation also appears to cause a westward intensification in the countercurrent. Larvae that become detrained from the Florida Current front will be transported westward and shoreward by the coastal countercurrent, providing ample opportunity for recruitment to the reefs and nearshore zones.

Because of the variable nature and mix of processes that form the recruitment conveyor system, larvae are provided with many opportunities for recruitment into the Tortugas and Florida Keys on time scales ranging from hours to months. Recruitment pathways providing even longer retention times, such as required by spiny lobster larvae, are also available, especially in the Tortugas region and are shown schematically on Fig. 31. This pathway requires movement onto the southwest Florida shelf and return to the Tortugas or Keys via either the Loop Current/Florida Current or by way of a mean south-southeastward flow on the west Florida shelf and through western Florida Bay, eventually entering the coastal countercurrent in the middle or lower Keys. The west Florida shelf/Loop Current pathway has been observed with satellite tracked drifters (Figs. 13 and 24) and can increase larval retention to eight months (Lee et al., 1994). The western Florida Bay/coastal countercurrent route is also quite plausible as shown by trajectories of surface drifters recently deployed west of Florida Bay (Figs. 21a, b), and observations of net southeastward flow through the tidal channels between the Keys (Smith, 1994).

A more recent example of the recruitment pathway connecting the nursery grounds of south Florida’s mangrove and seagrass habitats of the Everglades and Florida Bay with the pelagic ecosystems supporting larval development in the Florida Keys coastal waters is shown in Figure 33. Drifter 23113 was deployed in the Shark River discharge plume on Oct. 15, 1998. Surface currents transported the drifter to the southwest, reaching the Marquesas Keys on Nov. 1. The mean southwest current speed during this period was approximately 8 cm/s. Surprisingly, the drifter moved clockwise around Marquesas Key before it emerged into the coastal countercurrent and accelerated toward the Dry Tortugas at a mean speed of about 14 cm/s. Sea surface temperature patterns derived from satellite AVHRR images (Figs. 34a-f) shows that the coastal countercurrent was being forced by a system of Tortugas eddies that were moving eastward along the Florida Current front at this time. The drifter reached the vicinity of the Tortugas on Nov. 13 then converged with the Florida Current front on Nov. 15 and accelerated rapidly to the east at speeds of about 100 cm/s. The cyclonic circulation in a Tortugas eddy recirculated the drifter back into the coastal countercurrent in the vicinity of Long Key on Nov. 21 and 22 (Figs. 33, 34b and c), which once again returned the drifter to the Tortugas on Dec. 7 and 8 (Figs. 34c and d), where it was again entrained in the Florida Current front and accelerated to the east. The persistence of a series of Tortugas eddies offshore of the Keys during this period caused the drifter to recirculate with the eddies, extending as far north as Key Largo on Dec. 22 before it was once again caught by the countercurrent and transported westward along the outer fringing reefs (Figs. 33, 34d, e and f). At the time of this writing on Jan. 21, 1999 it was located near the Marquesas Keys and appeared poised to reenter the shallow nursery grounds of west Florida Bay after nearly three months of recirculating within the coastal eddies. This recirculating pattern of nearsurface currents is a common occurrence in the lower and western Keys and provides a conveyor system with many opportunities for larval recruitment into the Keys from both local and remote sources and may help to explain the high species diversity and large abundances in the region.

Seasonal changes in the conveyor system may also influence seasonal patterns in recruitment. The seasonal maximum in westward and southwestward alongshore winds in the fall can cause a seasonal maximum in the strength and northward extent of the coastal countercurrent, together with a maximum in onshore surface Ekman transports and a minimum in Florida Current downstream flow. These consequences should result in greater larval retention and enhanced opportunity for recruitment into the Tortugas and Keys coastal waters. During summer months winds are weak and have a northward component, and the Florida Current flow is maximum, causing a reduction in the spatial extent of the coastal countercurrent. The result could be a decrease in recruitment to the upper Keys as larvae are carried out of the coastal retention zone by the Florida Current, but recruitment in the lower Keys and Tortugas should continue with little interference.

The recruitment conveyor systems of the Florida Keys provide opportunity for larval retention and self recruitment to a wide assortment of local species with considerable differences in durations of their larval stages, as well as different behavior strategies and swimming capabilities. Table 3 lists larval durations for a selection of important species in the Keys. These pelagic stages range from a few hours for some of the reef building coral species up to a year for spiny lobster larvae. The highly variable nature of the size, strength, duration and translation speed of the eddies traveling along the Florida Current front, together with local tidal currents and wind driven currents provide a wide selection of retention times and recirculation patterns to match the wide range of larval duration periods. Eddies range in size from small-scale frontal instabilities with dimensions of 3-15 km that are carried downstream by the Florida Current at speeds of 30-50 cm/s and have short life-spans of 1-3 days (Lee, 1975; Lee and Mayer, 1977; Shay et al., 1998) to the large-scale Tortugas eddies, with dimensions of 100-200 km that move slowly downstream at speeds of 5-15 km/day and can persist up to 3 months (Lee et al., 1992; Lee et al., 1994; Fratantoni et al., 1998).

Self recruitment of locally spawned coral larvae with short pelagic stages of hours to days is aided by retention in small-scale frontal instabilities and spin-off eddies and by wind and tidal driven reversals in the coastal currents. The potential for local recruitment of the longer larval stages of some of the reef building corals and queen conch with planktonic stages of 2-4 weeks is enhanced by the passage of Tortugas eddies in the middle, lower and western Keys and also by the wind driven coastal counter current. The time-scale for recirculation within these eddies matches well with the larval duration periods. Pink shrimp are known to migrate from the Everglades/Florida Bay nursery grounds to the Dry Tortugas to spawn (Rehrer et al., 1967; Criales and Lee, 1995). This migration can benefit from the observed southwestward currents that occur in this region from east and northeast wind forcing (Figs. 21b, 23a and 33). After spawning in the Dry Tortugas, the recirculation of the Tortugas gyre provides a direct route for self recruitment of this fishery on the 3-4 week time-scale of the larvae (Criales and Lee, 1995). Rehrer et al. (1967) speculated that recruitment to the Everglades nursery region occurred through advection by the Florida Current front followed by movement through the Keys and western Florida Bay. Based on our more recent results this route also appears quite plausible as part of the conveyor system. Snapper/grouper are also known to spawn in the Tortugas and have a similar 3-4 week pelagic stage as the pink shrimp. Self recruitment of these species is also significantly enhanced by the recirculation of the Tortugas gyre. Self recruitment in the spiny lobster fishery presents the most serious challenge to larvae spawned locally in the Keys due to the long pelagic stages of 6-12 months. Previously, it was believed that most spiny lobster in the Keys were recruited from remote upstream sources in the Caribbean, and recent genetic studies do indeed reveal a homogeneity of the spiny lobster gene pool in the Caribbean, Bahamas and Florida waters (Silberman et al., 1994). However, genetic similarities do not elucidate the degree of isolation or connectivity of populations in different geographic regions. Abundances of local populations are determined more by the life history strategies of the animals, food supply and the oceanographic conditions affecting the larval, juvenile and adult stages. The dispersal and retention in ocean currents and eddies, together with advantageous animal behavior plays a major role in the size of populations in different regions by controlling the patterns and rates of recruitment (Roberts, 1997; Cowen and Castro, 1994; Cowen, 1998).

Recent oceanographic studies presented above have shown that there is a considerable potential for self recruitment in the lower and western Keys due to the combined influences of the Tortugas Gyre recirculation, coastal countercurrents, wind driven currents on the southwest Florida shelf and Loop Current/Florida Current boundary currents. Together these current systems form a larval conveyor system that can cause larval retention on the 6-12 month time-scale of the spiny lobster. In addition, it is well known that later stages of reef fish, shrimp and lobster larvae have considerable swimming ability allowing them to take advantage of horizontal and vertical current structures to prolong their retention or aid reentry into a juvenile habitat (Cowen and Castro, 1994). Even weak swimmers could adjust their position within the horizontal current reversals of the conveyor system or enhance onshore or offshore movements by vertical migration into either the onshore flow in the upper Ekman layer or the offshore flow within the lower Ekman layer, since these are persistent features in the downwelling regime of the lower and western Keys.

Summary

This work provides a description of oceanographic characteristics of the Dry Tortugas region using a synthesis of results from the literature, as well as recent and ongoing studies. Particular emphasis is placed on the influence of physical processes on larval recruitment from local and remote sources. The results presented are based primarily on the following recent and ongoing studies of the University of Miami: the South East Florida and Caribbean Recruitment study (SEFCAR); the South Florida Oil Spill Research Center study (SFOSRC); and the Florida Bay Circulation and Exchange Project of the South Florida Ecosystem Restoration Prediction and Modeling Program (SFERPM) study. Results of a completed MMS study of the physical oceanography of the Florida Current by SAIC were also of considerable use for describing the offshore conditions.

The findings presented above show clearly that the Dry Tortugas region is unique in its location and the extent to which oceanographic processes impact the area. But even more importantly, the Tortugas play a dynamic role in supporting marine ecosystems throughout south Florida and the Florida Keys. Larvae that are spawned from adult populations in the Tortugas are spread throughout the Keys and south Florida by a persistent system of currents and eddies that provide the retention and current pathways necessary for successful recruitment of both local and foreign spawned recruits with larval stages ranging from hours for some coral species up to one year for spiny lobster. In addition the upwellings and convergences of the current systems provide the necessary food supplies in concentrated frontal regions to support larval growth stages.

The Dry Tortugas are located at the transition between the Gulf of Mexico and the Atlantic. As such, they are strongly impacted by two major current systems, the Loop Current in the eastern Gulf of Mexico and the Florida Current in the Straits of Florida, as well as the system of eddies that form and travel along the boundary of these currents. Of particular importance to ecosystems of the Tortugas and Florida Keys is the formation of a large anticlockwise rotating gyre that forms just south of the Tortugas where the Loop Current turns abruptly into the Straits of Florida. This gyre can persist for several months before it is forced downstream along the Keys decreasing in size and increasing in forward speed until its demise in the middle Keys. The gyre serves as a retention mechanism for local recruits, and a pathway to inshore habitats for foreign recruits, as well as a potential food provider through plankton production and concentration.

The Dry Tortugas are also located adjacent to two coastal current systems, ie the wind driven currents of both the Florida Keys coastal zone and the west Florida shelf. Persistent westward winds of the Keys creates a downwelling system that drives a westward coastal countercurrent along the lower Keys to the Tortugas. The countercurrent provides a return route to the Tortugas and its gyre dominated circulation, and onshore surface Ekman transports provides a mechanism for larval entry into coastal habitats. Circulation on the west Florida shelf is strongly influenced by wind forcing but there also appears to occur a significant southward mean flow, possibly due to the Loop Current. The effect of these currents on the Tortugas is to provide a larval return mechanism to the Florida Bay nursery grounds during periods of southeast winds, as well as a transport mechanism for low-salinity shelf waters from the north when the mean southward flow is strong.

The location of the Tortugas adjacent to the west Florida shelf subjects the region to seasonal changes in vertical stratification. The west Florida shelf becomes strongly stratified in the late spring and summer due to the combined effects of atmospheric heating, increased river run-off and decrease wind mixing. During the late fall and winter the shelf becomes well mixed vertically from increased wind mixing, atmospheric cooling and decreased river run-off. These structures extend to the Tortugas, but do not extend to the Florida Keys due to isolation by the island chain, the shallow and narrow configuration of the coastal zone, and the strong influence of the Florida Current at the coastal boundary. The presence of summer stratification and vertical current shear in the Tortugas can be advantageous to recruitment of late stage larvae to the Evergaldes and Florida Bay nurseries, provided they possess suitable behavioral strategies to position themselves in the upper mixed layer where northward transports in the summer can occur.

The combination of downstream transport in the Florida Current, onshore Ekman transports along the downwelling coast, upstream flow in the coastal countercurrent and recirculation in the Tortugas gyre forms a recirculating recruitment pathway stretching from the Dry Tortugas to the middle Keys that enhances larval retention and recruitment into the Keys coastal waters of larvae spawned locally or foreign larvae from remote upstream areas of the Gulf of Mexico and Caribbean Sea. Convergences in the Florida Current front and coastal gyres provide a mechanism to concentrate foreign and local larvae, as well as their planktonic food supply. Onshore Ekman transports and horizontal mixing from frontal instabilities enhance export from the oceanic waters into the coastal zone. A wind- and gyre-driven coastal countercurrent provides a return leg to aid larval retention in local waters. Seasonal cycles of the winds, countercurrent and Florida Current favor recruitment to the coastal waters during the fall when the countercurrent can extend the length of the Keys from the Dry Tortugas to Key Largo, onshore Ekman transports are maximum and downstream flow in the Florida Current is minimum. The mix and variability of the different processes forming the recruitment conveyor provide ample opportunity for local recruitment of species with larval stages ranging from days to several months. For species with longer larval stages, such as the spiny lobster Panulirus argus, which has a 6 to 12 month larval period, a local recruitment pathway exists that utilizes retention in the Tortugas gyre and southwest Florida shelf and return via the Loop Current and the Keys conveyor system. Return from the southwest Florida shelf could also occur through western Florida Bay and the Keys coastal countercurrent, due to a net southeastward flow recently observed connecting the Gulf of Mexico to Atlantic through the Keys.

References

Bane, J. M., D. A. Brooks and K. R. Lorenson, 1981. Synoptic observations of the three-dimensional structure, propagation and evolution of Gulf Stream meanders along the Carolina continental margin. J. Geophys. Res., 86: 6411-6425.

Beardsley, R.C. and B. Butman, 1974. Circulation on the New England continental shelf: response to strong winter storms. Geophys. Res. Lett., 1: 181-184.

Bitterman, D. S. and Donald V. Hanson, The design of a low cost tropical drifter buoy. Proceedings MDS '86: Marine Data Systems International Symposium, 1986.

Brooks, D. A., and J. M. Bane, Jr., 1981. Gulf Stream fluctuations and meanders over the Onslow Bay upper continental slope. J. Phys. Oceanogr. 11: 247-256.

Brooks, I. H., and P. P. Niiler, 1975. The Florida Current at Key West: Summer 1972. J. Mar. Res., 33(1): 83-92.

Brooks, I., 1979. Fluctuations in the transport of the Florida Current at periods between tidal and two weeks. J. Phys. Oceanogr., 9: 1048-1053.

Chew, F., 1974. The turning process in meandering currents: A case study. J. Phys. Oceanogr., 4: 27-57.

Cowen, R. K. and L. R. Castro, 1994. Relation of coral reef fish larval distributions to island scale circulation around Barbados, West Indies. Bull. Mar. Sci., 54: 228-244.

Sale, P. F. and R. K. Cowen,1998. Comments Fishery and Reef Management. Science, 279: 2021.

Criales, M.M. and T.N. Lee. 1995. Larval distribution and transport of penaeoid shrimps during the presence of the Tortugas Gyre in May-June, 1991. U.S. Fisheries Bull. 93: 471-482.

Csanady, G. T., 1978. The arrested topographic wave. J. Phys. Oceanogr., 8: 47-62.

Csanady, G. T. 1982. Circulation in the Coastal Ocean. D. Reidel, Boston, Ma. 279 pp.

Duing, W., 1975. Synoptic studies of transients in the Florida Current. J. Mar. Res., 33, 53-73.

Duing, W., C. N. K. Mooers and T. N. Lee, 1977. Low-frequency variability in the Florida Current and relations to atmospheric forcing from 1972 to 1974. J. Mar. Res., 35, 129-161.

Fratantoni, P. S., 1998. The formation and evolution of Tortugas eddies in the southern Straits of Florida and Gulf of Mexico. A Dissertation, University of Miami, May 1998, 180 pg.

Fratantoni, P. S., T. N. Lee, G. P. Podesta and F. Muller-Karger, 1998. The influence of Loop Current perturbations on the formation and evolution of Tortugas eddies in the southern Straits of Florida. J. Geophys. Res., 103(C11): 24759-24799.

Fourqurean, J. W., R. D. Jones and J. C. Zieman, 1993. Processes influencing water column nutrient characteristics and phosphorus limitation of phytoplankton biomass in Florida Bay, USA; inferences from spatial distributions. Estuaries, Coastal and Shelf Science, 36: 295-314.

Gilbert, P. S., T. N. Lee and G. Podesta, 1996. Transport of anomalous low-salinity waters from the Mississippi river flood of 1993 to the Straits of Florida. Cont. Shelf Res. 16(8): 1065-1085.

Gordon, A. L., 1986. Interocean exchange of thermocline water. J. Geophys. Res., 91: 5037-5046.

Hurlburt, H. E., and J. D. Thompson, 1980. A numerical study of Loop Current intrusions and eddy shedding. J. Phys. Oceanogr., 10(10): 1611-1651.

Ichiye, T., H. Kuo and M. R. Carnes, 1973. Assessment of currents and hydrography of the eastern Gulf of Mexico, Control 601 Texas A & M University, College Park.

Johns, W. E. and F. Schott, 1987. Meandering and transport variations of the Florida Current. J. Phys. Oceanogr., 17: 1128-1147.

Leaman, K., R. Molinari and P. Vertes, 1987. Structure and variability of the Florida Current at 27oN: April 1982-July 1984. J. Phys. Oceanogr. 17: 565-583.

Lee, T. N., 1975. Florida Current spin-off eddies. Deep-sea Res., 22: 753-763.

Lee, T.N. 1986. Coastal circulation in the Key Largo Coral Reef Marine Sanctuary. Physics of Shallow Estuaries and Bays. In: Lecture Notes in Coastal and Estuarine Studies, Springer-Verlag. Vol. 16: 178-198.

Lee, T. N. and D. A. Mayer, 1977. Low-frequency current variability and spin-off Southeast Florida. J. Mar. Res., 35(1): 193-220.

Lee, T.N. and C.N.K. Mooers. 1977. Near-bottom temperature and current variability over the Miami slope and terrace. Bull. Mar. Sci. 27: 758-775.

Lee, T. N., and L. P. Atkinson, 1983. Low-frequency current and temperature variability from Gulf Stream frontal eddies and atmospheric forcing along the southeast U.S. outer continental shelf. J. Geophys. Res. 88: 4541-4568.

Lee, T.N., F. Schott and R. Zantopp, 1985. Florida Current: Low-frequency variability of the Florida Current as observed with moored current meters during April 1982-June 1983. Science, 227: 298-301.

Lee, T.N. and E. Williams. 1988. Wind forced transport fluctuations of the Florida Current. J. Phys. Oceanogr. 18: 937-946.

Lee, T. N., C. Rooth, E. Williams, M. McGowan, A. F. Szmant and M. E. Clarke. 1992. Influence of Florida Current, gyres and wind-driven circulation on transport of larvae and recruitment in the Florida Keys coral reefs. Cont. Shelf. Res., 12: 971-1002.

Lee, T. N., M. E. Clarke, E. Williams, A. F. Szmant and T. Berger, 1994. Evolution of the Tortugas gyre and its influence on recruitment in the Florida Keys. Bull. Mar. Sci., 54(3): 621-646.

Lee, T.N., K. Leaman, E. Williams, T. Berger and L. Atkinson. 1995. Florida Current meanders and gyre formation in the southern Straits of Florida. J. Geophys. Res., 100(C5): 8607-8620.

Lee, T. N., E. Johns, D. Wilson and E. Williams, 1998. Florida Bay circulation and exchange study. Proceedings of Florida Bay Science Conference, Miami Fl. May 12-14, 1998.

Leetmaa, A., P. P. Niiler and H. Stommel, 1977. Does the Sverdrup relation account for the mid-Atlantic circulation? J. Mar. Res., 35: 1-10.

Leipper, D. F., 1970. A sequence of current patterns in the Gulf of Mexico. J. Geophys. Res., 75(3): 637-657.

Lewis, J. B. 1951. The phyllosoma larvae of the spiny lobster Panulirus argus. Bull. Mar. Sci. Gulf Carib. 1: 89-103.

Limouzy-Paris, C.B., H.C. Graber, D.L. Jones, A. Ropke, and W.J. Richards, 1996: Translocation of larval coral reef fishes via sub-mesoscale spin-off eddies from the Florida Current. Bull. Mar. Sci., 60: 966-983.

Maul, G., 1977. The annual cycle of the Gulf Loop Current Part I: Observations during a one-year time series. J. Mar. Res. 35(1): 29-47.

Mitchum, G.T. and W. Sturges, 1982. Wind driven currents on the West Florida Shelf. J. Phys. Oceanogr., 12: 1310-1317.

Muller-Karger, F. E., J. J. Walsh, R. H. Evans, and M. B. Meyers, 1991. On the seasonal phytoplankton concentration and sea surface temperature cycles of the Gulf of Mexico as determined by satellites. J. Geophys. Res., 96(C7): 12645-12665.

Niiler, P. P. and W. S. Richardson. 1973. Seasonal variability in the Florida Current. J. Mar. Res. 31: 144-167.

Ortner, P., T. N. Lee, P. Milne, R. Zika, M. E. Clarke, G. P. Podesta, P. Swart, P. A. Tester, L. P. Atkinson and W. Johnson, 1995. Mississippi River flood waters that reached the Gulf Stream. J. Geophys. Res., 100(C7): 13595-13601.

Paluszkiewicz, T., L.P. Atkinson, E.S. Posmentier and C.R. McClain, 1983. Observations of a Loop Current Frontal Eddy intrusion onto the West Florida Shelf. J. Geophys. Res., 88: 9639-9652.

Pitts, P. A. 1994. An investigation of near-bottom flow patterns along and across Hawk Channel, Florida Keys. Bull. Mar. Sci. 54: 610-620.

Press, W. H., S. A. Teucholsky, W. T. Vetterling and B. P. Flannery. 1992. Numerical recipes in fortran: the art of scientific computing. Cambridge University Press, 2nd edition.

Rehrer, R., A. C. Jones and M. A. Roessler, 1967. Bottom water drift on the Tortugas Grounds. Bull. Mar. Sci., 17: 562-575.

Robblee, M. B., J. T. Tilmant and J. Emerson, 1989. Quantitative observations on salinity. Bull. Mar. Sci., 44: 523.

Roberts, C. M., 1997. Connectivity and management of Caribbean coral reefs. Science, 278: 1454-1457.

Rooth, C. and L. Xie. 1992. Air-sea boundary dynamics in the presence of mesoscale surface currents. J. Geophys. Res. 97: 14431-14438.

SAIC (Science Applications International Corporation), 1992. Straits of Florida physical oceanographic field study, Final Interpretive Report. Vol II, Rep. OCS Study 920024, 174 pp., Raliegh, N. C.

Schmitz, W. and P. L. Richardson, 1991. On the sources of the Florida Current. Deep-Sea Res., 38(Suppl.1): 5379-5409.

Schott, F., T. N. Lee and R. Zantopp, 1988. Variability of structure and transport of the Florida Current in the period range of days to seasonal. J. Phys. Oceangr., 18(9): 1209-1230.

Shay, L. K., T. N. Lee, E. Williams, H. Graber and C. Rooth, 1998. Effects of low-frequency current variability on near-inertial submesoscale vortices. J. Geophys. Res. 103: 18691-18714.

Silberman, J.D., S.K. Sarver and P.J. Walsh, 1994. Mitochondrial DNA variation and population structure in the spiny lobster, Panulirus argus. Mar. Biol. 120: 601-608.

Sims, H. W., Jr. and R. M. Ingle. 1966. Caribbean recruitment of Florida's spiny lobster population. Quarterly Journal of the Florida Academy of Science, 29: 207-243.

Smith, N. P. 1994. Long-term Gulf-to-Atlantic transport through tidal channels in the Florida Keys. Bull. Mar. Sci. 54: 602-609.

Stoner, A. W., N. Mehta, and T. N. Lee. 1997. Recruitment of Strombus Veligers to the Florida reef tract: relation to hydrographic events. J. Shellfish Res. 16: 1-6.

Sturges, W., 1992. An estimate of the spectrum of Loop Current variability from gappy data. J. Phys. Oceanogr., 22: 1245-1256.

Vukovich, F. M., 1988. On the formation of elongated cold perturbations off the Dry Tortugas. J. Phys. Oceanogr., 18: 1051-1059.

Vukovich, F. M. and G. A. Maul, 1985. Cyclonic eddies in the eastern Gulf of Mexico. J. Phys. Oceangr., 15: 105-117.

Vukovich, F. M. 1995. An updated evaluation of the Loop Currents eddy-shedding frequency. J. Geophys. Res., 100(5): 8655-8659.

Wang, J. D., 1998. Subtidal flow patterns in western Florida Bay. Estuaries, Coastal and Shelf Science, 46:901-915.

Wilson, W. D., and W. E. Johns, 1997. Velocity structure and transport in the Windward Island Passages. Deep-Sea Res., 44(3): 487-520.

Yeung, C. 1996. Larval transport and retention of lobster phyllosoma larvae in the Florida Keys, Ph. D. Dissertation. University of Miami, Coral Gables, 217 pp.

Yeung, C. and M. F. McGowan. 1991. Differences in inshore-offshore and vertical distribution of phyllosoma larvae of Panulirus, Scyllarus, and Scyllarides in the Florida Keys in May-June 1989. Bull. Mar. Sci. 49(3): 699-714.

Yeung, C., J. T. Couillard and M. F. McGowan. 1993. The relationship between vertical distribution of spiny lobster phyllosoma larvae (Crustacea: Palinuridae) and isolume depths generated by a computer model. Rev. Biol. Trop. 41(1): 63-67.

Zantopp, R., K.D. Leaman and T.N. Lee. 1987. Florida Current Meanders: A close look in June/July 1984. J. Phys. Oceanogr. 17: 584-595.