Abstract. We show the space and time variability of the Florida Current in the Southern Straits of Florida using imagery derived from measurements made with two high-frequency over-the-horizon (ionospheric) radars. Between March and August 1997, ten surface current maps show how the Florida Current meanders in response to mesoscale eddies propagating eastward through the Straits. Satellite altimetry aids in the interpretation of the radar imagery by showing upstream eddy development in the Gulf of Mexico.
________________________
1. NOAA Environmental Technology Laboratory, Boulder, CO 80303
2. Cooperative Institute for Research in Environmental Sciences, University of Colorado/ NOAA, Boulder, CO 80309
3. Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL 33149
4. Colorado Center for Astrodynamics Research, University of Colorado, Boulder, CO 80309
1. Introduction
Previously, we have shown how high-frequency (HF) over-the-horizon (OTH) radars can be used to map surface currents over very large ocean areas. We first demonstrated this ability using a single radar to map the radial component of surface current [Georges and Harlan, 1995; Georges et al., 1998], then by combining the measurements by two OTH radars to map vector surface currents [Georges et al., 1996]. Recently, we showed an example of surface-current mapping in the vicinity of an Atlantic hurricane [Harlan and Georges, 1998].
Although the information available from a single radar "snapshot" often reveals details of the spatial structure of ocean surface currents with new detail, a more complete picture of the dynamics of current systems requires a time sequence of such snapshots. To demonstrate this ability, we focused two OTH radars on the Southern Straits of Florida (SSF) between March and August 1997, in an attempt to obtain a sequence of measurements showing the dynamics of the Florida Current (FC).
2. Southern Straits of Florida
The Florida Straits have been the focus of many field and modeling studies concerned with (a) the coastal ecosystem and fisheries of the region [Lee et al., 1992], (b) the effect of mass and heat transported by the Western Boundary Current through the Straits on the general circulation of the North Atlantic Ocean [Leaman et al., 1987; Molinari, 1987], and (c) oil-spill risk assessment [Leaman et al., 1995; Atkinson et al., 1995]. An ongoing study of the South Florida coastal ecosystem seeks an understanding of the physical and biological processes responsible for the degradation of the coral reefs along the Florida Keys [Prospero and Ortner, 1995]. Such understanding requires measurements that quantify the interaction and exchange of Florida Bay waters with the connecting coastal waters of the Gulf of Mexico and with the Atlantic Ocean through the tidal passages between the Florida Keys. However, the physical mechanisms driving the flows through these passages are unknown, as is the connection to transient upstream forcing events such as variations in the Gulf of Mexico Loop Current and associated eddies.
Qualitative features of the Florida Current and subsidiary flows can be seen using sea-surface temperature (SST) imagery from satellites, such as the Advanced Very High Resolution Radiometer (AVHRR), but such data are unusable during the summer because of surface water heating and cloud cover. Acoustic current profilers and in-situ current meters have been deployed across critical passages in the straits to determine not only the main channel transport but also the contributions to the Florida Current by its subsidiary channels [Atkinson et al., 1995; Leaman et al., 1995], which account for about 10% of the mass transport of the Florida Current.
In all of these endeavors, the existing array of observing techniques would be usefully supplemented by a means for mapping the spatial distribution of surface currents and their evolution in time. To demonstrate what may be possible, we show here a sequence of surface current maps made with two military OTH radars that illuminated the SSF between March and August of 1997. Surface currents are, of course, only part of the four-dimensional picture of the circulation of any region of ocean. A complete picture is likely to emerge only by combining data from multiple sensors.
3. The Radars
In partnership with the US Navy Fleet Surveillance Support Command and the Raytheon Company, NOAA performs ocean-monitoring tests using two HF OTH radars known as ROTHR (Relocatable Over-the-Horizon Radar) in Texas and Virginia. Their coverage of the Caribbean Sea and portions of the Atlantic Ocean and Gulf of Mexico is achieved using 5-to-28-MHz radio waves that reflect from the ionosphere. Each ROTHR achieves a nominal 0.5 angular resolution with a 2.58-km-long linear phased receiving array consisting of 372 twin-monopole elements. Range resolution is achieved by transmitting a continuous frequency-modulated waveform. The ocean region studied in this test is a small fraction of the total coverage of the two radars, and coverage is at present limited by the amount of radar time we are given. Other examples of how we use these radars for oceanographic studies can be found on the NOAA website: http://www1.etl.noaa.gov/othr.
4. Background for HF Radar Measurements of Ocean Currents
The method for measuring near-surface ocean currents using high-frequency (HF: 3 to 30 MHz) radar was first demonstrated more than 20 years ago [Stewart and Joy, 1974; Barrick et al., 1977]. It is based on the fact that HF backscatter from the sea surface is primarily from Bragg-resonant surface waves, that is, those waves whose wavelength is exactly one-half the radio wavelength and which travel exactly toward and away from the radio-wave source. In the absence of underlying currents, deep-water ocean waves of a given wavelength always travel at the same phase velocity. Therefore, the spectrum of echoes from the sea surface at HF consists of two sharp "Bragg" lines, symmetrically displaced from zero Doppler, whose Doppler shifts correspond to that velocity. For example, a 15-MHz radar sees mainly ocean waves of 10-m wavelength, which travel at 4 m s-1 and produce echoes Doppler shifted by ± 0.4 Hz. An underlying surface current adds to or subtracts from the waves' apparent phase speed, displacing both Bragg lines from symmetry about zero Doppler. In the example, a 1 m s-1 advancing radial current shifts both lines by + 0.1 Hz. The radial component of surface current can thus be measured by tracking such small variations in the shift of the Bragg lines from symmetry about zero. Two radars illuminating a given ocean cell are required to measure surface current vectors.
Several shore-based HF current-mapping radar systems are commercially available, such as the Coastal Ocean Dynamics Radar (CODAR) [Lipa and Barrick, 1995] and the Ocean Surface Current Radar (OSCR) [Shay et al., 1995]. Because they employ surface-wave propagation, however, their range from shore is generally less than 50 km. HF over-the-horizon radars use exactly the same principle to map surface currents, except that they use ionospheric reflections to greatly extend the radar range, which can be up to 3500 km using one ionospheric "hop" [Trizna, 1982]. The area illuminated is thus increased from hundreds to millions of square kilometers. The price for this greatly extended coverage is the need for large-aperture transmitting and receiving antenna arrays and strategies to avoid and remove the distortion caused by ionospheric motions on the outgoing and returning radar paths [Georges et al., 1998]. These strategies include the use of nearby land echoes for zero-Doppler references and coordinate registration, an objective spectral quality (sharpness) filter, real-time sweep-frequency diagnostic soundings over the actual radar path to help select a radar frequency that is free of distorting multipath, and use of stable ionospheric layers that tend to occur near midday.
5. The Measurements
Between March and August 1997, we used both radars to interrogate a roughly rectangular 75,000 km2 region in the Southern Straits of Florida (SSF), about 1820 km from the Texas radar and about 1500 km from the Virginia radar (Fig. 1). During this time period, we obtained ten surface current maps of varying quality and at irregular intervals determined by radar access considerations.
Each radar mapped the distribution of radial surface currents, which are interpolated to a common grid and combined to form vector currents. The data were acquired in 25-s blocks, usually 10 to 15, spaced in time by 180 s, resulting in a total acquisition time of less than one hour, during which the radar is shared with other users. Radar frequency is variable and is selected for optimum sea-echo quality on each day using real-time sweep-frequency backscatter soundings of the ionosphere over the actual radar path. Frequencies as low as 11 MHz and as high as 19.5 MHz were used in this test, corresponding to Bragg-resonant ocean wavelengths between 13.6 and 7.7 m. For the waveform, frequencies and range used in this test, a radar resolution cell on the ocean surface is about 6 km in range and 10 km in azimuth. Based on validations conducted in the Gulf of Mexico using a shipborne acoustic Doppler current profiler (ADCP), we estimate the accuracy of the radar-derived current measurements to be 10-15 cm s-1 [Harlan et al., 1998].
6. Interpretation of the ROTHR Surface Current Fields
Dual ROTHR current data from the Southern Straits of Florida (SSF) provide a unique method for observing Florida Current (FC) variability on time scales greater than one week and spatial scales of 10 to 300 km. Recent studies of moored-current-meter records and satellite AVHRR imagery have shown that the dominant mode of FC variability in the SSF occurs on time scales of 30 - 70 days due to the movement and evolution of Tortugas eddies [Lee et al., 1995; Fratantoni et al., 1997]. These cold, cyclonic eddies form initially on the Loop Current front in the Gulf of Mexico, move into the SSF were they can remain stationary off the Dry Tortugas for up to three months and reach dimensions of 100 to 200 km before they are advected downstream, decreasing in size until their demise off the Middle Keys. The downstream advection in the SSF occurs at speeds of 5 to 15 km/day and causes a southward deflection or meander of the FC axis that travels downstream in phase with the eddy [Lee et al., 1992; 1995; Fratantoni et al., 1997].
The ROTHR-derived current fields clearly resolve the surface current response to the dominant eddy/meander mode of the FC in the SSF (Figs. 2a-2j). These data also show the presence of significant currents in the shallow waters north of the Florida Keys just west of Florida Bay and the occurrence of an anti-cyclonic eddy north of the central Cuban coast (Fig. 2g). Additional evidence supporting the existence and evolution of the features shown in the ROTHR surface current data are provided by satellite AVHRR imagery (not shown but in agreement with the location, shape and dimensions of the FC features in the ROTHR data before June, when sea surface temperature becomes nearly uniform and clouds are more prevalent), and by sea surface height fields derived from satellite altimetry. The sea surface height (SSH) fields are a blend of Topex/Poseidon and ERS-2 sea-surface-height anomalies plus a model mean sea level. Sea surface height fields are available daily from the website of the Colorado Center for Astrodynamics Research and provide a comparison with the ROTHR current fields in the SSF but also show the evolution of features in the Gulf of Mexico upstream of the SSF. The SSH fields are shown in Figures 3a-3p for the same days as the ROTHR data and for times in between the ROTHR fields when they were more than one week apart.
7. Comparison of SSH with ROTHR current fields
From March 19 to May 24, the SSH fields show the Loop Current (LC) extending northward into the Gulf from about 25N to near 26.5N and interacting with a large cyclonic eddy northwest of the LC (Figs. 3a-3e). During this period a Tortugas eddy is located on the north side of the FC near 82.5W and propagates out of the SSH domain near 81W at a speed of about 5 km/day by May 24. The ROTHR current fields show that the effect of the Tortugas eddy is to displace the FC axis southward toward the coast of Cuba on March 19 and April 30 (Fig. 2a and 2b).
On May 14, the ROTHR domain was shifted to the east and indicates that the offshore-displaced FC is interacting with the shallow Cay Sal Bank that is centered near 24N and 80W (Fig. 2c), causing a branch of southward flow around the west side of the bank and into the Santaren Channel. There is also a well-resolved cyclonic eddy in the Santaren Channel between Cay Sal and Great Bahamas Banks. This feature was previously observed with moored current data from this same location during a similar southward meander of the FC [Lee et al., 1995]. The eastward flow indicated on the Great Bahamas Bank east of 79 W is most likely tidal.
During the period June 5 to July 7, the SSH fields show a large southward meander of the FC near 83W that developed into a strong Tortugas eddy by July 7 (Figs. 3f-k). The deflection of the FC around the eddy appears to bring the current axis very near the north coast of Cuba in this region. From July 7 to August 7, the eddy propagated eastward through the SSF at a mean speed of about 5 km/day and was located between 81W and 82W on August 7 (Figs. 3k-p). Interestingly, the intensification of the Tortugas eddy and its subsequent eastward movement occurred during a period of intense interaction between the LC and a cyclonic frontal eddy on its eastern border between 25N and 26N that appears to be the direct cause of a ring separation by July 29 (Figs. 3k-p). ROTHR current fields during this period clearly show the effect of the southward meander of the FC off the Dry Tortugas, causing the FC axis to be displaced close to the north coast of Cuba between 83W and 82W before turning northeastward and converging toward the Florida Keys (Figs. 2d and 2e).
The northward turning of the strong currents in the FC axis becomes more pronounced off the coast of Cuba on July 7 and 14 (Figs. 2f and 2g) associated with the intensification of the Tortugas eddy, which steers FC flow northward around its eastern side. The eastward movement of the Tortugas eddy through the SSF is clearly shown in the ROTHR current fields on July 29 and August 4 and 7 from the eastward displacement of the FC offshore meander (Figs. 2h-2j). On August 7, the eddy extends about 130 km along the Keys from the Marquesas to Marathon and about 50 km offshore in the vicinity of Big Pine Key. Recirculation and the formation of a countercurrent is evident on the northern side of the eddy. The location, shape, dimensions and movement of the eddy as determined from the ROTHR and SSH data sets are generally in close agreement.
In addition to FC variability, the ROTHR derived currents also indicate the presence at times of a clockwise eddy off the north coast of Cuba between 82W and 80W where the FC separates from the coast (Figs. 2f and 2g). The appearance of this feature could also be aided by a westward wind-driven coastal current along the north shore. In the shallow (less than 20 m) shelf waters north of the Keys, between Key West and Naples, the ROTHR currents show onshore (eastward) and offshore (westward) flows on the order of 30 to 40 cm s-1 (Figs. 2d-2j). The magnitude and direction of these flows agree well with moored current measurements from the region and are primarily tidal driven.
8. Conclusions
Pictures of the mesoscale variability of ocean surface currents provided by OTH radar are intended to complement the views provided by satellite and in-situ sensors. OTH radar looks at fixed ocean areas on demand, whereas satellites cover the globe in swaths dictated by orbital dynamics and sensor field of view. Furthermore, only large-scale geostrophic currents are derivable from SSH data. In-situ sensors are usually deployed during programs of limited space and time scope, but they sample the ocean depths in detail not obtainable by other means. Each of these instruments measures different ocean properties with varying reliability and resolution. Further studies using merged products should lead to better understanding of the four-dimensional dynamics of ocean features of practical relevance. In the immediate future, OTH radar studies will focus on the dynamics of the Gulf of Mexico Loop Current, its associated eddies, and the flow through the Yucatan Channel.
Acknowledgments
We thank the Office of the Secretary of Defense and the US Navy for granting access to their over-the-horizon radar system, and the Raytheon Company, and T. Dotson, in particular, for acquiring the ROTHR data.
References
Atkinson, L. P., T. Berger, P. Hamilton, E. Weddell, K. D. Leaman and T. N. Lee, Current meter observations in the Old Bahama Channel, J. Geophys. Res. 100(C5), 8555-8560, 1995.
Barrick, D. E., M. W. Evans, and B. Weber, Ocean surface currents mapped by radar, Science 198, 138-144, 1977.
Frantantoni, P., T. N. Lee, G. P. Podesta, and F. Muller-Karger, The influence of Loop Current perturbations on the formation and evolution of Tortugas eddies in the southern Straits of Florida, submitted to J. Geophys. Res., 1997.
Georges, T. M. and J. A. Harlan, Gulf Stream anomaly seen by defense radar, Eos, Trans. Amer. Geophys. Union 76(15), 146, 1995.
Georges, T. M, J. A. Harlan, R. R. Leben, and R. A. Lematta, A test of ocean surface current mapping with over-the-horizon radar, IEEE Trans. Geosci. Remote Sens. <January, 1998>.
Georges, T. M., J. A. Harlan and R. A. Lematta, Large-scale mapping of ocean surface currents with dual over-the-horizon radars, Nature 379, 434-436, 1996.
Harlan, J. A. and T. M. Georges, Observations of hurricane Hortense with two over-the-horizon radars, Geophys. Res. Lett., in press.
Harlan, J. A., T. M. Georges and D. C. Biggs, Comparison of over-the-horizon radar surface current measurements in the Gulf of Mexico with simultaneous sea truth, J. Marine Sys., in press.
Leaman, K.D., R.L. Molinari, and P.S. Vertes, Structure and variability of the Florida Current at 27°N:April 1982-July 1984. J. Phys. Oceanogr., 17, 565-583, 1987.
Leaman, K. D., P. S. Vertes, L. P. Atkinson, T. N. Lee, P. Hamilton and E. Waddell, Transport, potential vorticity, and current/temperature structure across Northwest Providence and Santaren Channels and the Florida Current off Cay Sal Bank, J. Geophys. Res. 100(C5), 8561-8569, 1995.
Lee, T.N., C. Rooth, E. Williams, M. McGowan, A.F. Szmant and M.E. Clarke, Influence of Florida Current, gyres and wind-driven circulation on larvae transport and recruitment in the Florida Keys Coral Reefs., Cont. Shelf Res. 12(78), 971-1002, 1992.
Lee, T.N., K. Leaman, E. Williams, T. Berger and L. Atkinson, Florida Current meanders and gyre formation in the Southern Straits of Florida, J. Geophys. Res. Oceans 100, 8607-8620, 1995.
Lipa, B. J. and D. E. Barrick, Networking multiple HF radar systems with common coverage overlap, Proc. IEEE Fifth Conf. on Current Measurements, 7-9 Feb., 1995, IEEE Pub.95CH, 1995.
Molinari, R.L., A review of STACS results on Florida Current variability, Eos, Transactions of the Amer. Geophys. Union, 68(16), 337, 1987.
Prospero, J.M., and P.B. Ortner, South Florida Coastal Ocean Ecosystem Workshop, Miami, FL, February 27-28, 1995., University of Miami, Cooperative Institute for Marine and Atmospheric Studies, 2 vols., 1995.
Shay, L. K., H. C. Graber, D. B. Ross, and R. D. Chapman, Mesoscale ocean surface current structure detected by high-frequency radar, J. Atmos. Oceanic. Technol. 12, 881-900, 1995.
Stewart, R. H. and J. W. Joy, HF radio measurements of ocean surface currents, Deep Sea Res. 21, 1039-1049, 1974.
Trizna, D. B., Mapping ocean currents using over-the-horizon HF radar, Int. J. Remote Sens. 3(3), 295-309, 1982.
FIGURE CAPTIONS
Figure 1. Geography of the Florida Straits showing the area illuminated by two over-the-horizon radars between March and August 1997.
Figures 2a through 2j. Surface current maps in the Southern Straits of Florida, produced by two over-the-horizon radars in Texas and Virginia between March and August 1997.
Figure 3 (a through f). Gulf of Mexico sea-surface height, derived from blended TOPEX/Poseidon and ERS-2 altimetry and a model mean.
Figure 3 (g through l). Gulf of Mexico sea-surface height, derived from blended TOPEX/Poseidon and ERS-2 altimetry and a model mean.
Figure 3 (m through p). Gulf of Mexico sea-surface height, derived from blended TOPEX/Poseidon and ERS-2 altimetry plus a model mean.