• Abascal, A. J., , J. Sheinbaum, , J. Candela, , J. Ochoa, , and A. Badan, 2003: Analysis of flow variability in the Yucatan Channel. J. Geophys. Res., 108, 3381, doi:10.1029/2003JC001922.

    • Search Google Scholar
    • Export Citation
  • Athie, G., , J. Candela, , J. Ochoa, , and J. Sheinbaum, 2012: Impact of Caribbean cyclones on the detachment of Loop Current anticyclones. J. Geophys. Res., 117, C03018, doi:10.1029/2011JC007090.

    • Search Google Scholar
    • Export Citation
  • Auer, S. J., 1987: Five-year climatological survey of the Gulf Stream system and its associated rings. J. Geophys. Res., 92 (C11), 11 70911 726.

    • Search Google Scholar
    • Export Citation
  • Barnett, T. P., 1983: Interaction of the monsoon and Pacific trade wind system at interannual time scale: Part I. Mon. Wea. Rev., 111, 756773.

    • Search Google Scholar
    • Export Citation
  • Barnier, B., , L. Siefridt, , and P. Marchesiello, 1995: Thermal forcing for a global ocean circulation model from a three-year climatology of ECMWF analysis. J. Mar. Syst., 6, 363380.

    • Search Google Scholar
    • Export Citation
  • Batchelder, H., 2006: Forward-in-time-/backward-in-time-trajectory (FITT/BITT) modeling of particles and organisms in the coastal ocean. J. Atmos. Oceanic Technol., 23, 727741.

    • Search Google Scholar
    • Export Citation
  • Candela, J., , J. Sheinbaum, , J. Ochoa, , A. Badan, , and R. Leben, 2002: The potential vorticity flux through the Yucatan Channel and the Loop Current in the Gulf of Mexico. Geophys. Res. Lett., 29, 2059, doi:10.1029/2002GL015587.

    • Search Google Scholar
    • Export Citation
  • Candela, J., , S. Tanahara, , M. Crepon, , B. Barnier, , and J. Sheinbaum, 2003: Yucatan Channel flow: Observations versus CLIPPER ATL6 and MERCATOR PAM models. J. Geophys. Res., 108, 3385, doi:10.1029/2003JC001961.

    • Search Google Scholar
    • Export Citation
  • Carton, J. A., , and Y. Chao, 1999: Caribbean Sea eddies inferred from TOPEX/Poseidon altimetry and a ⅙° Atlantic Ocean model simulation. J. Geophys. Res., 104 (C4), 77437752.

    • Search Google Scholar
    • Export Citation
  • Chao, Y., , and M. S. Lozier, 2001: Evaluation of North Atlantic property field simulations at ⅙°. J. Phys. Oceanogr., 31, 32003213.

    • Search Google Scholar
    • Export Citation
  • Chao, Y., , A. Gangopadhyay, , F. Bryan, , and W. R. Holland, 1996: Modeling the Gulf Stream system: How far from reality? Geophys. Res. Lett., 23, 31553158.

    • Search Google Scholar
    • Export Citation
  • Chérubin, L. M., , Y. Morel, , and E. P. Chassignet, 2006: Loop Current ring shedding: The formation of cyclones and the effect of topography. J. Phys. Oceanogr., 36, 569591.

    • Search Google Scholar
    • Export Citation
  • Cochrane, J. D., 1972: Separation of an anticyclone and subsequent developments in the Loop Current. Contributions on the Physical Oceanography of the Gulf of Mexico, L. R. A. Capuro and J. L. Reid, Eds., Gulf Publishing, 91–106.

    • Search Google Scholar
    • Export Citation
  • Ducet, N., , P. Y. LeTraon, , and G. Reverdin, 2000: Global high-resolution mapping of ocean circulation from TOPEX/Poseidon and ERS-1 and -2. J. Geophys. Res., 105, 477498.

    • Search Google Scholar
    • Export Citation
  • Dukowicz, J. K., , and R. D. Smith, 1994: Implicit free-surface method for the Bryan–Cox–Semtner ocean model. J. Geophys. Res., 99 (C4), 79918014.

    • Search Google Scholar
    • Export Citation
  • Elliott, B. A., 1982: Anticyclonic rings in the Gulf of Mexico. J. Phys. Oceanogr., 12, 12921309.

  • 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), 24 75924 779.

    • Search Google Scholar
    • Export Citation
  • Hofmann, E. E., , K. S. Hedstrom, , J. R. Moisan, , D. B. Haidvogel, , and D. L. Mackas, 1991: Use of simulated drifter tracks to investigate general transport patterns and residence times in the coastal transition zone. J. Geophys. Res., 96 (C8), 15 04115 052.

    • Search Google Scholar
    • Export Citation
  • Horel, J. D., 1984: Complex principal component analysis: Theory and examples. J. Climate Appl. Meteor., 23, 16601673.

  • Hurlburt, H. E., 1986: Dynamic transfer of simulated altimeter data into subsurface information by a numerical ocean model. J. Geophys. Res., 91 (C2), 23722400.

    • Search Google Scholar
    • Export Citation
  • Hurlburt, H. E., , and J. D. Thompson, 1982: The dynamics of Loop Current and shed eddies in a numerical model of the Gulf of Mexico. Hydrodynamics of Semi-Enclosed Seas, J. C. J. Nihoul, Ed., Elsevier Scientific, 243–297.

    • Search Google Scholar
    • Export Citation
  • Leben, R., , G. Born, , and B. Engebrecht, 2002: Operational altimeter data processing for mesoscale monitoring. Mar. Geod., 25, 318.

  • Leipper, D. F., 1970: A sequence of current patterns in the Gulf of Mexico. J. Geophys. Res., 75, 637657.

  • LeTraon, P. Y., , Y. Faugère, , F. Hernandez, , F. M. J. Dorandeu, , and M. Ablain, 2003: Can we merge Geosat follow-on with TOPEX/Poseidon and ERS-2 for an improved description of the ocean circulation? J. Atmos. Oceanic Technol., 20, 889895.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., , and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 117 pp.

  • Levitus, S., , R. Burgett, , and T. P. Boyer, 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 99 pp.

  • Lugo-Fernandez, A., 2007: Is the Loop Current a chaotic oscillator? J. Phys. Oceanogr., 37, 14551469.

  • Matsuura, T., , and T. Yamagata, 1982: On the evolution of nonlinear planetary eddies larger than the radius of deformation. J. Phys. Oceanogr., 12, 440456.

    • Search Google Scholar
    • Export Citation
  • Maul, G. A., 1977: The annual cycle of the Gulf Loop Current. Part I: Observations during a one-year time series. J. Mar. Res., 35, 2947.

    • Search Google Scholar
    • Export Citation
  • Maul, G. A., , and F. M. Vukovich, 1993: The relationship between variations in the Gulf of Mexico Loop Current and Straits of Florida volume transport. J. Phys. Oceanogr., 23, 785796.

    • Search Google Scholar
    • Export Citation
  • Merino, M., 1997: Upwelling on the Yucatan Shelf: Hydrographic evidence. J. Mar. Syst., 13, 101121.

  • Molinari, R. L., , J. F. Festa, , and D. W. Behringer, 1978: The circulation in the Gulf of Mexico derived from estimated dynamic height fields. J. Phys. Oceanogr., 8, 987996.

    • Search Google Scholar
    • Export Citation
  • Murphy, S. J., , H. E. Hurlbert, , and J. J. O'Brien, 1999: The connectivity of eddy variability in the Caribbean Sea, the Gulf of Mexico, and the Atlantic Ocean. J. Geophys. Res., 104 (C1), 14311453.

    • Search Google Scholar
    • Export Citation
  • Nakamura, M., , and Y. Chao, 2001: Diagnoses of an eddy-resolving Atlantic Ocean model simulation in the vicinity of the Gulf Stream. Part I: Potential vorticity. J. Phys. Oceanogr., 31, 353378.

    • Search Google Scholar
    • Export Citation
  • Ochoa, J., , J. Sheinbaum, , A. Baden, , J. Candela, , and D. Wilson, 2001: Geostrophy via potential vorticity inversion in the Yucatan Channel. J. Mar. Res., 59, 725747.

    • Search Google Scholar
    • Export Citation
  • Oey, L.-Y., , H.-C. Lee, , and W. J. Schmitz Jr., 2003: Effects of winds and Caribbean eddies on the frequency of Loop Current eddy shedding: A numerical model study. J. Geophys. Res., 108, 3324, doi:10.1029/2002JC001698.

    • Search Google Scholar
    • Export Citation
  • 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, 96399652.

    • Search Google Scholar
    • Export Citation
  • Pichevin, T., , and D. Nof, 1997: The momentum imbalance paradox. Tellus, 49, 298319.

  • Pullin, D. I., 1978: The large-scale structure of unsteady self-similar rolled-up vortex sheets. J. Fluid Mech., 88, 401430.

  • Pullin, D. I., , and A. E. Perry, 1980: Some flow visualization experiments on the starting vortex. J. Fluid Mech., 97, 239255.

  • Rio, M. H., , S. Guinehut, , and G. Larnicol, 2011: New CNES-CLS09 global mean dynamic topography computed from the combination of GRACE data, altimetry, and in situ measurements. J. Geophys. Res., 116, C07018, doi:10.1029/2010JC006505.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, L. K., 1985: Introduction to the CODE-2 moored array and large-scale data report. CODE-2 Moored Array and Large-Scale Data Rep., R. Limeburner, Ed., CODE Tech. Rep. 38, WHOI Tech. Rep. 85–35, 21 pp.

  • Rott, N., 1956: Diffraction of a weak shock with vortex generation. J. Fluid Mech., 1, 111128.

  • Schmitz, W. J., Jr., , and M. S. McCartney, 1993: On the North Atlantic circulation. Rev. Geophys., 31, 2949.

  • Sheinbaum, J., , J. Candela, , A. Baden, , and J. Ochoa, 2002: Flow structure and transports in the Yucatan Channel. Geophys. Res. Lett., 29, 1040, doi:10.1029/2001GL013990.

    • Search Google Scholar
    • Export Citation
  • Signell, R. P., , and W. R. Geyer, 1991: Transient eddy formation around headlands. J. Geophys. Res., 96 (C2), 25612575.

  • Stern, M. E., , and G. R. Flierl, 1987: On the interaction of a vortex with a shear flow. J. Geophys. Res., 92 (C10), 10 73310 744.

  • Sturges, W., 1992: An estimate of the spectrum of Loop Current variability from gappy data. J. Phys. Oceanogr., 22, 12451256.

  • Sturges, W., , and J. C. Evans, 1983: On the variability of the Loop Current in the Gulf of Mexico. J. Mar. Res., 41, 639653.

  • Sturges, W., , and R. Leben, 2000: Frequency of ring separation from the Loop Current in the Gulf of Mexico: A revised estimate. J. Phys. Oceanogr., 30, 18141819.

    • Search Google Scholar
    • Export Citation
  • Sturges, W., , N. G. Hoffmann, , and R. R. Leben, 2010: A trigger mechanism for Loop Current ring separations. J. Phys. Oceanogr., 40, 900913.

    • Search Google Scholar
    • Export Citation
  • van Senden, D. C., , and J. Imberger, 1990: Effects of initial conditions and Ekman suction on tidal outflows from inlets. J. Geophys. Res., 95 (C8), 13 37313 391.

    • Search Google Scholar
    • Export Citation
  • Vukovich, F. M., 1988: On the formation of elongated cold perturbation off the Dry Tortugas. J. Phys. Oceanogr., 18, 10511059.

  • Vukovich, F. M., 1995: An updated evaluation of the Loop Current eddy-shedding frequency. J. Geophys. Res., 100 (C5), 86558659.

  • Vukovich, F. M., , and G. A. Maul, 1985: Cyclonic eddies in the eastern Gulf of Mexico. J. Phys. Oceanogr., 15, 105117.

  • Walker, N. D., , S. Myint, , A. Babin, , and A. Haag, 2003: Advances in satellite radiometry for the surveillance of surface temperatures, ocean eddies and upwelling processes in the Gulf of Mexico using GOES-8 measurements during summer. Geophys. Res. Lett., 30, 1854, doi:10.1029/2003GL017555.

    • Search Google Scholar
    • Export Citation
  • Walker, N. D., , R. R. Leben, , S. Anderson, , J. Feeney, , P. Coholan, , and N. Sharma, 2009: Study of Loop Current frontal eddies based on satellite remote sensing and drifter data. U.S. Department of the Interior, Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA, OCS Study MMS 2009-023, 78 pp.

  • Walker, N. D., and Coauthors, 2011: Impacts of a Loop Current frontal eddy cyclone and wind forcing on the 2010 Gulf of Mexico oil spill. Monitoring and Modeling the Deepwater Horizon Oil Spill: A Record-Breaking Enterprise, Geophys. Monogr., Vol. 195, Amer. Geophys. Union, 103–116.

  • Wallace, J. M., , and R. E. Dickinson, 1972: Empirical orthogonal representation of time series in the frequency domain: I. Theoretical considerations. J. Appl. Meteor., 11, 887892.

    • Search Google Scholar
    • Export Citation
  • Wells, M. G., , and G.-J. F. van Heijst, 2003: A model of tidal flushing of an estuary by dipole formation. Dyn. Atmos. Oceans, 37, 223244.

    • Search Google Scholar
    • Export Citation
  • Zavala-Hidalgo, J., , S. L. Morey, , and J. J. O'Brien, 2003: Cyclonic eddies northeast of the Campeche Bank from altimetry data. J. Phys. Oceanogr., 33, 623629.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Bottom topography (m) and schematic position of the LC boundary (thick line) in the Gulf of Mexico. Various geographic locations in the Gulf of Mexico are indicated. The triangle represents the Dry Tortugas. Dashed line A indicates the cross section in which time series of eddy kinetic energy (EKE) and volume transport are shown in Fig. 5. Box B indicates the region in which volume-averaged relative vorticity is calculated in Fig. 8. Dashed–dotted line C delineates the cross section where temperature and relative vorticity at 104.5 m are shown in Fig. 10. Box D delineates the region where CEOF analysis is performed.

  • View in gallery

    Snapshots of temperature (°C) and velocity vectors at 104.5 m in the ⅙° Atlantic Ocean general circulation numerical model. Solid red lines represent isobaths (200 and 1000 m). Eddies labeled W and X are Caribbean anticyclones going toward the Yucatan Channel. The eddy labeled G is an LCFE moving along the boundary of the LC. Yellow points represent particles initially seeded in eddy W on day 678 and their subsequent locations. Pink points are particles seeded in LCFE G on day 744 and their previous locations using a backward-in-time tracking method. Day 1 represents the first day in the 5-yr model output.

  • View in gallery

    Temperature (°C) and meridional velocity (cm s−1) sections in the ⅙° Atlantic Ocean numerical model on day 702 along 23.05°N: contour intervals are 1°C and 10 cm s−1; positive value (solid line) is northward velocity and negative value (shaded region and dashed line) is southward.

  • View in gallery

    As in Fig. 3, but for along the 25.82°N section on day 726.

  • View in gallery

    Time series of (a) area-averaged EKE of the normal velocity component across the YC (section A in Fig. 1), (b) total volume transport into the GOM, (c) volume transport above 800 m into the GOM, and (d) volume transport below 800 m into the GOM for the 5-yr output period of the ⅙° Atlantic Ocean general circulation model. Also indicated in (a) are events of a Caribbean anticyclonic eddy approaching the YC (marked by arrows), which occur just before the time of local EKE peak.

  • View in gallery

    (a) Power spectrum of the area-averaged EKE of the normal velocity component across the YC and (b) power spectra of total volume transport into the GOM (dashed line), volume transport above 800 m (boldface line) and below 800 m (thin line) into the GOM.

  • View in gallery

    Time series of the LC areal extension as calculated from the area encompassed by the 23°C isotherm at 104.5 m in the GOM, LC warm ring–shedding events (marked by black dots, numbers indicate days), and events of Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows) for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model.

  • View in gallery

    Time series of (a) volume-averaged (box B in Fig. 1) cyclonic (thin line) and negative anticyclonic (boldface line) relative vorticity, (b) low-pass-filtered (filter half-amplitude period 100 days) cyclonic relative vorticity in (a), and (c) high-pass-filtered cyclonic relative vorticity [series in (a) minus series in (b)] for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated in (c) are events of Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows), which occur just before the time of local relative vorticity peak.

  • View in gallery

    Power spectrum of the volume-averaged cyclonic relative vorticity shown in Fig. 8a.

  • View in gallery

    Time–latitude plots of (a) temperature (°C) and (b) relative vorticity (s−1) along the Campeche Bank continental slope (section C in Fig. 1) at 104.5-m depth for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated are events of a Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows), which occur concomitantly with events of cold water penetration and enhanced cyclonic vorticity over the Campeche Bank slope. Two examples of LCFE propagation along this section are indicated by the dashed–dotted lines.

  • View in gallery

    Spatial structure of the first two CEOFs of the high-pass-filtered (filter half-amplitude period 100 days) temperature anomaly (box D in Fig. 1) at 104.5 m for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model: (a) amplitude and (b) phase (°) of the first CEOF mode. (c),(d) As in (a),(b), but for the second CEOF mode. Boldface lines represent isobaths (200 and 1000 m) in the region.

  • View in gallery

    Temporal evolution of principal components of the first two CEOFs (Fig. 11): (a) amplitude and (b) phase (°) of the first complex principal component. (c),(d) As in (a),(b), but for the second complex principal component.

  • View in gallery

    Time–latitude plots of (a) temperature (°C) reconstructed from the second CEOF mode and (b) high-pass-filtered (filter half-amplitude period 100 days) temperature (°C) anomalies from Fig. 10a along the Campeche Bank continental slope (section C in Fig. 1) at 104.5-m depth for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated are events of Caribbean anticyclonic eddy approaching the YC (marked by arrows). The two dashed–dotted lines indicate LCFE propagation along this section.

  • View in gallery

    Near-real-time SSH (cm) pattern in the GOM and the Caribbean Sea on (a) 15 Feb 2002 showing an anticyclone near the YC and a weak cyclone along the western margin of the LC just north of the YC, and (b) 27 Feb 2002 showing the anticyclone within the YC and an intensified cyclone north of the anticyclone.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 6 6 3
PDF Downloads 4 4 4

An Analysis of Loop Current Frontal Eddies in a ⅙° Atlantic Ocean Model Simulation

View More View Less
  • 1 Department of Oceanography and Coastal Sciences, School of the Coast and Environment, Louisiana State University, Baton Rouge, Louisiana
  • 2 Department of Earth, Ocean and Atmospheric Science, The Florida State University, Tallahassee, Florida
  • 3 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California
  • 4 Colorado Center for Astrodynamics Research, University of Colorado Boulder, Boulder, Colorado
© Get Permissions
Full access

Abstract

The Loop Current frontal eddies (LCFEs) refer to cyclonic cold eddies moving downstream along the outside edge of the Loop Current in the eastern Gulf of Mexico. They have been observed by in situ measurements and satellite imagery, mostly downstream of the Campeche Bank continental shelf. Their evolution, simulated by a primitive equation ⅙° and 37-level Atlantic Ocean general circulation numerical model, is described in detail in this study. Some of the simulated LCFEs arise, with the passage through the Yucatan Channel of a Caribbean anticyclonic eddy, as weak cyclones with diameters less than 100 km near the Yucatan Channel. They then grow to fully developed eddies with diameters on the order of 150–200 km while moving along the Loop Current edge. Modeled LCFEs have a very coherent vertical structure with isotherm doming seen from 50- to ~1000-m depth. The Caribbean anticyclone and LCFE are two predominant features in this numerical model simulation, which account for 22% and 10%, respectively, of the short-term (period less than 100 days) temperature variance at 104.5 m in the complex empirical orthogonal function (CEOF) analysis. The source water inside the LCFEs that are generated by Caribbean anticyclonic eddy impingement can be traced back, using a backward-in-time Lagrangian particle-tracking method, to the western edge of the Caribbean Current in the northwest Caribbean Sea and to coastal waters near the northern Yucatan Peninsula. The model results indicating a pairing of anticyclonic and cyclonic eddies within and north of the Yucatan Channel are supported by satellite altimetry measurements during February 2002 when several altimeters were operational.

Current affiliation: Remote Sensing Solutions, Pasadena, California.

Corresponding author address: Haosheng Huang, 320 Howe-Russell Geoscience Complex, Department of Oceanography and Coastal Sciences, School of the Coast and Environment, Louisiana State University, Baton Rouge, LA 70803. E-mail: hhuang7@lsu.edu

Abstract

The Loop Current frontal eddies (LCFEs) refer to cyclonic cold eddies moving downstream along the outside edge of the Loop Current in the eastern Gulf of Mexico. They have been observed by in situ measurements and satellite imagery, mostly downstream of the Campeche Bank continental shelf. Their evolution, simulated by a primitive equation ⅙° and 37-level Atlantic Ocean general circulation numerical model, is described in detail in this study. Some of the simulated LCFEs arise, with the passage through the Yucatan Channel of a Caribbean anticyclonic eddy, as weak cyclones with diameters less than 100 km near the Yucatan Channel. They then grow to fully developed eddies with diameters on the order of 150–200 km while moving along the Loop Current edge. Modeled LCFEs have a very coherent vertical structure with isotherm doming seen from 50- to ~1000-m depth. The Caribbean anticyclone and LCFE are two predominant features in this numerical model simulation, which account for 22% and 10%, respectively, of the short-term (period less than 100 days) temperature variance at 104.5 m in the complex empirical orthogonal function (CEOF) analysis. The source water inside the LCFEs that are generated by Caribbean anticyclonic eddy impingement can be traced back, using a backward-in-time Lagrangian particle-tracking method, to the western edge of the Caribbean Current in the northwest Caribbean Sea and to coastal waters near the northern Yucatan Peninsula. The model results indicating a pairing of anticyclonic and cyclonic eddies within and north of the Yucatan Channel are supported by satellite altimetry measurements during February 2002 when several altimeters were operational.

Current affiliation: Remote Sensing Solutions, Pasadena, California.

Corresponding author address: Haosheng Huang, 320 Howe-Russell Geoscience Complex, Department of Oceanography and Coastal Sciences, School of the Coast and Environment, Louisiana State University, Baton Rouge, LA 70803. E-mail: hhuang7@lsu.edu

1. Introduction

The circulation in the eastern Gulf of Mexico (GOM) is dominated by the Loop Current (LC) flow. The LC originates from the Yucatan Current, an inflow of warm Caribbean water through the Yucatan Channel (YC). The flow then loops anticyclonically in the eastern GOM, forming the LC and exits through the Straits of Florida where it becomes the Florida Current. The LC path migrates northward into the GOM to varying degrees over the course of approximately a year and eventually retreats to the south leaving behind an anticyclonic warm ring that propagates into the western GOM. This path variability of the LC has been the focus of many studies (e.g., Maul 1977; Molinari et al. 1978; Sturges and Evans 1983; Vukovich 1988; Pichevin and Nof 1997; Walker et al. 2003; Sturges et al. 2010). The anticyclonic warm rings generated by the LC have a diameter of 150–400 km and move westward at 2.1–6.2 km day−1 (Elliott 1982; Auer 1987). The ring separation periods are very irregular and vary from 6 to 17 months (Sturges 1992; Maul and Vukovich 1993; Vukovich 1995; Sturges and Leben 2000; Oey et al. 2003; Lugo-Fernandez 2007; Walker et al. 2009).

In addition to the anticyclonic warm ring, observations have also revealed a wide range of smaller-scale cyclonic cold features that are superimposed on the LC. Leipper (1970) investigated current patterns in the northeastern GOM and found meanders and eddies located along the northern and eastern edges of the LC. This led to the first notion that anticyclonic ring shedding might be preceded by the westward propagation of cyclonic eddies across the LC (Cochrane 1972). Paluszkiewicz et al. (1983) found a cold-core frontal eddy similar to those observed on the southeastern U.S. shelf, which was carried southward by the LC along the seaward edge of the West Florida Shelf. Based on more than 10 years of satellite infrared data and several years of coincidental hydrographic measurements, Vukovich and Maul (1985) identified large cold-water perturbations that moved southward along the LC boundary adjacent to the West Florida Shelf. These perturbations eventually grew into stationary meanders with closed streamlines and cold cores near Dry Tortugas. Diameters of these features varied from 100 to 200 km with subsurface signatures detectable from near the surface to ~1000 m. Once the perturbations reached Dry Tortugas, they were observed to either dissipate or grow westward across the width of the LC. Fratantoni et al. (1998) and Walker et al. (2009) showed that the Tortugas cold eddies evolved from cyclonic frontal eddies that formed along the edge of the LC and moved downstream with the LC proper. Walker et al. (2011) showed that eddy merging processes can lead to significantly larger cyclonic eddies on the northern margin of the LC with important consequences to the offshore transport of river water and surface contaminants such as oil. To follow the convention in the literature, in this paper cyclonic eddies are called Loop Current frontal eddies (LCFEs) if they are along the western, northern, and eastern edges of the LC and Tortugas eddies if they are near Dry Tortugas.

Despite the aforementioned studies, however, the origin of LCFEs is still poorly understood. For example, Vukovich and Maul (1985) pointed out that cold-water perturbations were first observed along the LC northwestern boundary where the water depth was more than 1000 m. But, it was not clear whether they were developed locally or might have existed upstream. Zavala-Hidalgo et al. (2003) identified eight cyclonic eddies near the western edge of the LC, at the shelf break northeast of the Campeche Bank, through Ocean Topography Experiment (TOPEX)/Poseidon sea surface height (SSH) anomaly data. They suggested that the formation of the cyclones might be related to the dynamics of the LC because their generation appeared to be time locked with the warm-ring-shedding events. Nevertheless, no information was given regarding the specific mechanism of LCFE formation.

Numerical ocean circulation models are instrumental in exploring possible eddy generation mechanisms. Hurlburt (1986) gave evidence in his model experiments that LCFEs were continuously formed owing to baroclinic instabilities near the steep topography of the Campeche Bank. Chérubin et al. (2006), from an analysis of the MICOM simulation, showed that cyclones were products of vortex rim instability.

Extensive direct measurements of flow structure through the YC were carried out by the Canek program (Ochoa et al. 2001; Sheinbaum et al. 2002; Candela et al. 2002, 2003; Abascal et al. 2003). Although not intended for LCFE observations, these measurements casted new light into detailed velocity structures and transport variations upstream of the LCFE area, which were considered by many researchers as principal drivers of the LC and related mesoscale variability. Further analysis of the data revealed a strong relationship between the potential vorticity flux through the YC and the evolution of the LC in the GOM (Candela et al. 2002). Their analysis suggested that variability of the LC was probably correlated with the northward passage of anticyclones and cyclones from the Caribbean (Abascal et al. 2003). Recently, Athie et al. (2012) suggested that cyclonic vorticity anomalies generated in the western Cayman Sea could propagate northward along Caribbean coast of Mexico and cause an eastward shift of the Yucatan Current core just before several anticyclonic-eddy-shedding events in the GOM.

In the present study, LCFEs are observed in the output of a ⅙° and 37-level Atlantic Ocean general circulation numerical model (Chao et al. 1996). Some of these cold frontal eddies appear to be triggered by the passage of Caribbean anticyclonic eddies through the YC. The eddy activity in the Caribbean Sea in the same numerical model has been documented by Carton and Chao (1999). They found that large cyclonic and anticyclonic eddies appeared frequently in good agreement with TOPEX/Poseidon altimeter observations. Murphy et al. (1999) also found anticyclonic eddies in the Caribbean Sea in their multilayer numerical simulations. Their study showed that the anticyclonic eddies traversed the Caribbean Basin from west of the Lesser Antilles island arc to the YC and sometimes passed through the YC into the GOM. Presumably a certain response must be triggered in the LC when a Caribbean anticyclone passes through the YC. The purpose of this paper is to document such a response as revealed from a numerical general circulation model.

The paper is organized as follows. In section 2, a brief introduction of the Atlantic Ocean general circulation model is given. Section 3 presents a description of the life cycle of an LCFE in the model result. In section 4, various time series analyses and a complex empirical orthogonal function (CEOF) analysis on the ⅙° numerical model output are performed to further examine plausible linkages between the Caribbean anticyclones and LCFEs. A discussion is given in section 5. Conclusions are drawn in section 6.

2. Introduction of the Atlantic Ocean general circulation model

The Atlantic Ocean general circulation numerical model, in which LCFEs are reproduced, has been described in detail in Chao et al. (1996), Carton and Chao (1999), and Chao and Lozier (2001). Briefly, it was based on the Parallel Ocean Program (POP) developed at the Los Alamos National Laboratory (Dukowicz and Smith 1994). The model domain covered the Atlantic basin from 35°S to 80°N, 100°W to 20°E. The water exchange processes across the artificially closed boundaries were parameterized by 5°-wide buffer zones in which the model temperature and salinity were restored toward the hydrographic climatology (Levitus et al. 1994; Levitus and Boyer 1994). The horizontal grid spacing was approximately ⅙° (0.1843° latitude by 0.1875° longitude). There were 37 vertical levels with 19 levels concentrated in the upper 1000 m. The model was forced by the climatological monthly mean wind stress and heat flux derived from the European Centre for Medium-Range Weather Forecasts (ECMWF) analysis during 1983–86 (Barnier et al. 1995). Sea surface salinity was restored to the Levitus et al. (1994) and Levitus and Boyer (1994) climatology with a time scale of 30 days.

The model has been integrated for a total of 40 years starting from a null velocity field and the climatological January temperature and salinity for each level. Model output during the last 5 years was used in the present work. Due to limited data storage, the full three-dimensional model output was available at 3-day intervals. The model flow and property fields agreed reasonably well with observations of the North Atlantic Ocean, especially in the upper ocean (Chao et al. 1996; Carton and Chao 1999; Nakamura and Chao 2001; Chao and Lozier 2001). In this study, focus is placed on the Gulf of Mexico and northwest corner of the Caribbean Sea that covers 18°–31°N and 98°–80°W (Fig. 1).

Fig. 1.
Fig. 1.

Bottom topography (m) and schematic position of the LC boundary (thick line) in the Gulf of Mexico. Various geographic locations in the Gulf of Mexico are indicated. The triangle represents the Dry Tortugas. Dashed line A indicates the cross section in which time series of eddy kinetic energy (EKE) and volume transport are shown in Fig. 5. Box B indicates the region in which volume-averaged relative vorticity is calculated in Fig. 8. Dashed–dotted line C delineates the cross section where temperature and relative vorticity at 104.5 m are shown in Fig. 10. Box D delineates the region where CEOF analysis is performed.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

In the model, the LC path appeared reasonably well simulated in both temperature and velocity fields. The LC mean volume transport across the YC was 22.7 Sv (1 Sv ≡ 106 m3 s−1) during the 5-yr period, ranging from a minimum of 12.6 Sv to a maximum of 32.6 Sv. There were 13 LC warm rings detached from the LC proper in the course of 5 years. Three of them reattached to the LC afterward. The time between two consecutive events of warm-ring shedding could be as short as 3 months or as long as 13 months. Although the model-simulated LC volume transport is somewhat lower than the nominally accepted transport value [e.g., Schmitz and McCartney (1993) gave an estimation of 30 Sv mean transport through the Straits of Florida], it is close to the direct transport measurement of 23.8 ± 1 Sv in the Yucatan Channel from the Canek program (Ochoa et al. 2001; Sheinbaum et al. 2002; Candela et al. 2003). The LC warm-ring detachment frequency is also consistent with satellite observation between 2006 and 2009 (Athie et al. 2012).

3. A case study of the evolution of an LCFE in the ⅙° numerical model

Figure 2 presents one case in which mesoscale cyclonic frontal eddies were found moving along the boundary of the LC in the numerical model. Since summer insolation makes the sea surface temperature (SST) uniformly high in most of the GOM, temperature distribution and current vectors at 104.5-m depth were used to depict circulation patterns. The LCFE manifested itself as a cold counterclockwise rotating eddy at this depth.

Fig. 2.
Fig. 2.

Snapshots of temperature (°C) and velocity vectors at 104.5 m in the ⅙° Atlantic Ocean general circulation numerical model. Solid red lines represent isobaths (200 and 1000 m). Eddies labeled W and X are Caribbean anticyclones going toward the Yucatan Channel. The eddy labeled G is an LCFE moving along the boundary of the LC. Yellow points represent particles initially seeded in eddy W on day 678 and their subsequent locations. Pink points are particles seeded in LCFE G on day 744 and their previous locations using a backward-in-time tracking method. Day 1 represents the first day in the 5-yr model output.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Although the numerical model was forced by the ECMWF climatological monthly mean wind stress and surface heat flux, the generation of LCFEs in the model exhibited little seasonality. Therefore, no explicit dates are mentioned in the discussion below. The day numbers presented below use the first day in the 5-yr time series as day 1.

The most interesting feature in Fig. 2 is the presence of an LCFE on the western boundary of the LC and the simultaneous passage of a Caribbean anticyclonic eddy through the YC. In Fig. 2a, the LC penetrated north to ~25°N, and a recently separated LC warm ring was seen centered at 25°N, 90°W. An anticyclonic eddy, labeled W in the figure, was moving northward along the eastern side of the LC. It had a circular shape and was more than 300 km in diameter. In Fig. 2b part of anticyclone W was in the YC, just to the east side of the LC. It changed to an elliptic shape, with a long axis along the direction of the continental slope in that area. A cold patch (labeled G in the figure), with a dimension of about 100 km and core temperature less than 18°C, was found on the cyclonic shear side of the LC where the water depth ranges from 200 to 1000 m. The southward velocity on the western side of the cold-core eddy G was about 30 cm s−1. In the next few days, the Caribbean anticyclone W squeezed through the YC and coalesced with the LC proper. In the meantime, the cold-core eddy G intensified and moved northward along the LC western boundary over the Campeche Bank continental slope. On day 714 (Fig. 2c), eddy G left the continental slope and was partly in the deep GOM region. There was an LC warm filament wrapping around it from the north. Its movement after this point is consistent with the description of LCFEs from early studies (e.g., Vukovich and Maul 1985; Fratantoni et al. 1998). The dimension of eddy G increased to approximately 150–200 km on day 726 (Fig. 2d). The LC proper had an enhanced anticyclonic circulation owing to the addition of the Caribbean warm eddy W.

In Fig. 2e, eddy G approached the west Florida continental slope. It stayed there for about 18 days with little size and position change. During this time period, another Caribbean anticyclone X impinged on the Yucatan Peninsula, squeezed through the YC, and coalesced with the anticyclonically circulating LC proper (Figs. 2e and 2f). An upwelling-induced cold feature was almost always seen adjacent to the northeast tip of the Campeche Bank slope in Fig. 2. The upwelling intensified as anticyclone X moved through the YC (Fig. 2f). However, no obvious southward flow was found off the cyclonic side of the LC near the Campeche Bank. Therefore, no LCFE was generated by Caribbean anticyclone X.

From day 762 onward, the LCFE G near the West Florida Shelf began to move southward along the eastern boundary of the LC. At the same time, its core temperature increased and the counterclockwise circulating velocity weakened. In Fig. 2h it was found near Dry Tortugas, in the western entrance of the Straits of Florida.

A Lagrangian particle-tracking model, based on simulated velocity vectors at 104.5-m depth, was used to compute water parcel horizontal movement in this region. Velocities needed to compute the particle paths were interpolated from the model velocity field using a trilinear interpolation scheme in longitude, latitude, and time. The particle positions were integrated through time using a fourth-order Runge–Kutta algorithm (Hofmann et al. 1991). To test particle-tracking accuracy, 15 particles were seeded in a circle centered in the Caribbean anticyclone W on day 678 (Fig. 2a, yellow dots). They were observed to follow the trajectory of anticyclone W, for example, moving across the YC, merging with the LC proper, and being part of LC water on day 744 (Fig. 2e). This gives us some confidence in our particle-tracking method, although the 3-day time interval of model output may be a limiting factor. To detect where the LCFE water came from, a backward-in-time tracking method was used (Batchelder 2006) in which particle positions were identified at earlier times from their known destinations. The computer code for backward-in-time tracking was almost the same as that for forward-in-time tracking, with the sole difference in that current velocity directions were reversed. Another group of 15 particles were seeded in an ellipse around the LCFE G on day 744 (Fig. 2e, pink dots). They were seen to follow the trajectory of cyclone G backward in time. On day 678 (Fig. 2a) part of the particles were located near the northern coast of the Yucatan Peninsula in the GOM and the other part on the western edge of the Caribbean Current near the northwest corner of the Caribbean Sea.

Figure 3 shows a vertical section through the cold eddy center at 23.05°N on day 702 when LCFE G was located over the continental slope of the Campeche Bank (same day as in Fig. 2b). The model revealed that the signature of this cold eddy was quite coherent throughout the whole water column. The doming of isotherms could be traced from near the sea surface to the ocean bottom at depths 600–800 m. The frontal eddy was quite strong at this time, with maximum southward velocity reaching 40 cm s−1 while the maximum LC velocity was ~185 cm s−1. When this LCFE moved into the deeper Gulf of Mexico where water depth exceeds 2000 m, the doming of isotherms could be detected even below 1000 m (Fig. 4). The almost vertical configuration of northward and southward flowing zones in Fig. 4 indicates that the LCFE had a quite coherent vertical structure in the ⅙° numerical simulation. This assertion was further confirmed by the horizontal temperature and velocity fields at 1125 m in which the LCFE still manifested itself as a relatively cold patch with counterclockwise velocity (figure not shown).

Fig. 3.
Fig. 3.

Temperature (°C) and meridional velocity (cm s−1) sections in the ⅙° Atlantic Ocean numerical model on day 702 along 23.05°N: contour intervals are 1°C and 10 cm s−1; positive value (solid line) is northward velocity and negative value (shaded region and dashed line) is southward.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Fig. 4.
Fig. 4.

As in Fig. 3, but for along the 25.82°N section on day 726.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

4. Time series analysis of the ⅙° Atlantic Ocean model simulation result

The LCFE generation hypothesis, that some of the LCFEs are initiated by interaction of Caribbean anticyclonic eddies with the LC near the YC, was further examined by performing various time series analyses using outputs from the ⅙° Atlantic Ocean model. Caution should be taken in interpreting analysis outcomes since 3-day sampling intervals were used in the numerical model output without invoking any low-pass filtering. Nonetheless, aliasing was not considered a serious problem in this case because climatological monthly mean wind stress and heat flux were used to force the model, and sea surface salinity was also restored to the Levitus climatology with a 30-day time scale. Previous experiences with ocean general circulation models indicate that the spectral structure in the numerical ocean model output is simpler than in the real ocean and model response to low-frequency forcing tends to concentrate mostly in the low-frequency band.

a. Caribbean anticyclonic eddy entrance in the Yucatan Channel

The normal velocity to a zonal section at 21.4°N (section A in Fig. 1) in the YC is decomposed into the sum of a mean flow and its fluctuations, as
e1
The sectional integral of is the net volume transport into the GOM, while the area average of the sectional integral of represents the area-averaged EKE of the velocity component normal to section A (hereafter referred to as EKE in the YC). Figure 5 shows the time series of EKE (Fig. 5a), total volume transport (Fig. 5b), transport above 800 m (Fig. 5c), and below 800 m (Fig. 5d) across the section. The most prominent feature in Fig. 5a is the existence of a dominant oscillation with period of about 70 days, whose amplitude is on the order of 0.02–0.04 m2 s−2. The moments just before the peak EKE values correspond well with arrival time of Caribbean anticyclonic eddies at section A as indicated by arrows in Fig. 5a, which were determined by visual inspection of horizontal temperature contours and velocity vectors at 104.5-m depth. The EKE spectrum [using five Hanning filters with bandwidth 0.003 cycles per day (cpd) and 11.4 degrees of freedom] further revealed that there was only one major strong and narrow peak centered at 71-day (Fig. 6a). Therefore, the EKE in the YC is a good indication of northward propagation of Caribbean anticyclones. There were ~25 anticyclonic eddies passing through the strait in the 5-yr period. There are a few exceptions to the aforementioned correlation, especially the one near day 471 when the EKE peak corresponded with a locally generated cyclonic eddy, instead of a Caribbean anticyclone, moving across section A in the horizontal fields.
Fig. 5.
Fig. 5.

Time series of (a) area-averaged EKE of the normal velocity component across the YC (section A in Fig. 1), (b) total volume transport into the GOM, (c) volume transport above 800 m into the GOM, and (d) volume transport below 800 m into the GOM for the 5-yr output period of the ⅙° Atlantic Ocean general circulation model. Also indicated in (a) are events of a Caribbean anticyclonic eddy approaching the YC (marked by arrows), which occur just before the time of local EKE peak.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Fig. 6.
Fig. 6.

(a) Power spectrum of the area-averaged EKE of the normal velocity component across the YC and (b) power spectra of total volume transport into the GOM (dashed line), volume transport above 800 m (boldface line) and below 800 m (thin line) into the GOM.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

The volume transport below 800 m (Fig. 5d) was characterized by fluctuations of various frequencies with no obvious peaks (Fig. 6b) while that above 800 m (Fig. 5c), not surprisingly, exhibited a very similar structure to the total transport (Fig. 5b). Spectral analysis showed that the latter two records had similar spectral structure. Two peaks are shown in Fig. 6b, significant at the 95% confidence level. On the high-frequency portion, of particular interest was the occurrence of 71-day peak that had the exact same frequency as in the EKE spectrum and, thus, seemed to be related to the influence of Caribbean anticyclone passage on the volume transport in the YC. The low-frequency fluctuation was centered at day 234. As will be shown later, this frequency fell in the low-frequency band appearing in the spectra of the LC areal extension record and relative vorticity series in the LC region, hence, reflects a Yucatan transport variation that is related to the large-scale LC dynamics.

b. Loop Current eddy-shedding events and Caribbean anticyclonic eddies

It is of special interest to know if any correlation exists between LC eddy-shedding events and that of Caribbean anticyclonic eddies squeezing through the YC. The dynamics of the LC can be, to some degree, represented by the areal extension of the LC water in the GOM. To avoid possible ambiguity between surface temperature of the LC water and that of ambient waters in summer months, the area bounded by the 23°C isotherm at 104.5 m in the GOM, which is traditionally considered the LC water around this depth (Maul 1977), was used to monitor variability of the LC proper. The time series is shown in Fig. 7 (solid line), in which oscillations with periods longer than 300 days are dominant. Spectral analysis (figure not shown) reveals that almost all of the energy is in a broad low-frequency band (period longer than 180 days) with the only peak at day 374. Also illustrated in Fig. 7 are the approximate time instances when warm rings were detected, from visual inspection of temperature and velocity vector time series at 104.5 m, to separate from the LC proper. It can be seen that during the 5-yr simulation period there were 10 warm rings detached (if a ring was detached, reattached, and redetached, then the two detachments were counted once and the first detachment time was used). Oftentimes LC warm-ring-shedding events occurred when the Loop Current extension area was near its local maximum. Ring separations also occurred when the LC penetration into the GOM was minimal. For these events, the size of the detached warm ring was relatively small (e.g., events occurred around day 636, day 717, and day 1446). Comparing Caribbean anticyclone propagation through the YC (depicted by the arrows in Fig. 7) with LC dynamics, it appears that there is no obvious relationship between the timing of warm-core-ring shedding and the timing of Caribbean anticyclone passage through the YC.

Fig. 7.
Fig. 7.

Time series of the LC areal extension as calculated from the area encompassed by the 23°C isotherm at 104.5 m in the GOM, LC warm ring–shedding events (marked by black dots, numbers indicate days), and events of Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows) for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Another index that can be used to monitor the LC dynamics is the volume-averaged relative vorticity in the LC region. Accompanying the northward penetration or southward retreat of the LC front in the eastern GOM, the relative vorticity is also expected to increase or decrease at almost the same cycle. Box B in Fig. 1 indicates the horizontal extension of the volume chosen for the vorticity calculation. The integration is performed inside this rectangle and from the sea surface to the ocean bottom. Sensitivity tests indicate that the conclusion derived below is not very sensitive to the horizontal size of the box, and the time evolution of relative vorticity for the whole water column is similar to that for the upper 800-m water column.

The time series of volume-averaged cyclonic and minus anticyclonic vorticity in box B are shown in Fig. 8a. The two sets of series were highly correlated (correlation coefficient 0.84 < 0.86 < 0.87, 95% confidence interval). This indicates that the relative vorticity in this area is mainly controlled by the large-scale LC dynamics and related warm-ring events, which have the largest contribution to the relative vorticity budget. The cyclonic and anticyclonic vorticity had almost the same magnitude and oscillated nearly in phase. The spectrum of the cyclonic vorticity series revealed that its fluctuation had two major components (Fig. 9). The low-frequency components (period longer than 150 days) possessed most of the energy and were centered around two peaks at day 220 and day 374. This variation was obviously related to the LC northward penetration events in the eastern GOM as inferred from linear regression analysis (maximum correlation coefficient was 0.79 at a lag of 36 days) between the low-pass-filtered [Rosenfeld (1985) filter with a half-amplitude period of 100 days] cyclonic vorticity series (Fig. 8b) and the areal extension of the LC (Fig. 7). Although the high-frequency peak in the spectrum of the cyclonic vorticity series had less energy, it was still significant at the 95% confidence interval (the 95% confidence interval for the spectrum at period 72 days is ~0.4 × 10−10). Its period of 72 days was very close to the 71-day peak in EKE series in the YC. The high-pass-filtered cyclonic vorticity time series [the original series (Fig. 8a) minus the low-pass-filtered series (Fig. 8b)] is shown in Fig. 8c. For most peaks in this record, there are preceding Caribbean anticyclonic eddy events in the YC, as illustrated by arrows in the figure. This suggests the following scenario of possible impact of Caribbean anticyclone passage on the large-scale LC. When an anticyclonic eddy impinges upon the Yucatan Current near the YC, it increases the velocity of the current. This will, subsequently, increase both cyclonic and anticyclonic vorticity of the Loop Current inside the GOM. This phenomenon is mostly independent of the LC warm-ring-shedding event because there is a clear gap in the spectral representation of the two events (Fig. 9) and they can be nicely separated in the time domain as shown in Fig. 8.

Fig. 8.
Fig. 8.

Time series of (a) volume-averaged (box B in Fig. 1) cyclonic (thin line) and negative anticyclonic (boldface line) relative vorticity, (b) low-pass-filtered (filter half-amplitude period 100 days) cyclonic relative vorticity in (a), and (c) high-pass-filtered cyclonic relative vorticity [series in (a) minus series in (b)] for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated in (c) are events of Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows), which occur just before the time of local relative vorticity peak.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Fig. 9.
Fig. 9.

Power spectrum of the volume-averaged cyclonic relative vorticity shown in Fig. 8a.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

c. Caribbean anticyclones and LCFEs

Even though an LCFE can occur when the LC flows in a relatively straight path between the Yucatan Channel and the Straits of Florida with little northward penetration, most of them are seen to propagate over the eastern continental slope of the Campeche Bank in the numerical model. The temperature and relative vorticity distribution at 104.5 m along track C (see Fig. 1) on the Campeche Bank slope are plotted in Fig. 10. Figure 10a shows that cold temperature near the YC is a persistent feature in this numerical simulation, which is also consistent with observed upwelling in this area (Merino 1997; Walker et al. 2003). However, accompanying the passage of every Caribbean anticyclonic eddy through the YC (indicated by the arrows in Fig. 10), a strong cold event can almost always be found over the Campeche Bank continental slope (Fig. 10a) with enhanced cyclonic vorticity near the YC (Fig. 10b). Most of these cold features do not travel along the slope and usually dissipate south of 24°N. Nonetheless, some of them have longer lives and are seen to move across 24°N latitude where the water depth plunges to beyond 1000 m [e.g., the cold-core events occurred at day 354 and day 1569 (Fig. 10a, dashed–dotted lines)]. Figure 10b shows that these cold features have cyclonic relative vorticity, therefore, they are identified as LCFEs. It should be reminded that some LCFEs do not propagate over the Campeche Bank. Even when they do so, due to the misalignment of eddy trajectories and track C, the Hovmöller diagrams may not clearly capture their signals by cutting through the eddy centers. The LCFE shown in Fig. 2 is such an example because its representation in temperature and vorticity fields (around day 702 in Fig. 10) are both very weak compared to other eddies.

Fig. 10.
Fig. 10.

Time–latitude plots of (a) temperature (°C) and (b) relative vorticity (s−1) along the Campeche Bank continental slope (section C in Fig. 1) at 104.5-m depth for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated are events of a Caribbean anticyclonic eddy approaching the Yucatan Channel (marked by arrows), which occur concomitantly with events of cold water penetration and enhanced cyclonic vorticity over the Campeche Bank slope. Two examples of LCFE propagation along this section are indicated by the dashed–dotted lines.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

To get a better understanding of the spatial structure and temporal variability of the LCFE generation process in the numerical simulation, a CEOF analysis (Wallace and Dickinson 1972; Barnett 1983) of the temperature data at 104.5 m was performed. The CEOF analysis works best in identifying propagating disturbances and is more accurate when the energy of the time series is confined in a narrow-frequency band (Horel 1984). The spectral structures of the LC variation and the Caribbean anticyclonic eddy impingement events illustrated before suggest two disparate frequency peaks, one with a period longer than 150 days and the other centered around 70 days. Therefore, the temperature data at 104.5 m in a rectangular box (box D in Fig. 1) during the 5-yr model simulation period was first high-pass filtered with a half-amplitude period of 100 days (Rosenfeld 1985). Then, the series at every spatial point was normalized by the corresponding standard deviations from which a large state matrix was constructed. The CEOF procedure described by Barnett (1983) was used to obtain complex eigenvectors (i.e., CEOF) of the correlation matrix and their corresponding complex principal components, whose amplitudes and phases allowed the detection of moving patterns.

Figure 11 displays amplitude and phase functions of the spatial component of the first two CEOF modes, which accounted for 22% and 10% of the variance of the temperature field respectively. Temporal evolutions of these modes are shown in Fig. 12. The spatial structure of the first mode, with maximum variability in the Caribbean Sea and near the YC (Fig. 11a), suggests that this mode is likely related to the passages of Caribbean anticyclones through the YC. The CEOF phase associated with this mode (Fig. 11b) increases in a nearly linear fashion in both the Caribbean and Gulf of Mexico a small distance away from the YC, while near the YC section the phase change is relatively rapid. The first-order interpretation is that temperature anomalies approach the GOM from the Caribbean Sea and their amplitudes decrease as they travel toward the Straits of Florida. The rapid phase change near the YC may reflect the size and shape deformation of the warm eddies when they squeeze through the narrow strait. The spectra of the real and imaginary part of the first principal component (figure not shown) are almost identical and both reveal a single narrow peak centered at day 71, which is exactly the frequency of Caribbean anticyclone impingement upon the LC seen in the EKE series (Fig. 6a). From this evidence, it is suggested that the first mode of the temperature CEOF analysis at 104.5 m is dominated by passages of Caribbean anticyclonic eddies through the YC.

Fig. 11.
Fig. 11.

Spatial structure of the first two CEOFs of the high-pass-filtered (filter half-amplitude period 100 days) temperature anomaly (box D in Fig. 1) at 104.5 m for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model: (a) amplitude and (b) phase (°) of the first CEOF mode. (c),(d) As in (a),(b), but for the second CEOF mode. Boldface lines represent isobaths (200 and 1000 m) in the region.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

Fig. 12.
Fig. 12.

Temporal evolution of principal components of the first two CEOFs (Fig. 11): (a) amplitude and (b) phase (°) of the first complex principal component. (c),(d) As in (a),(b), but for the second complex principal component.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

The spatial amplitude and relative phase for the second CEOF mode (Figs. 11c and 11d) show that the largest variability in this mode occurs over the continental slope of the Campeche Bank, off the Campeche Bank in the deep GOM, just to the west of the West Florida Shelf near 26°N, and at the entrance of the Straits of Florida. The configuration of perturbation centers looks very similar to the LCFEs rotating around the edge of the LC, which has penetrated to about 25.5°N. The phase relation indicates that these perturbations propagate in the same direction as LCFEs and Tortugas eddies. The phase of this principal component (Figs. 12d) shows that the oscillations are more irregular than in the first mode, with full cycle oscillations (see zero crossings on the phase plot) lasting as long as 100 days or as short as 20 days. The energy in the real and imaginary part of the principal component of this mode (figure not shown) has a single peak at 66 days, but its bandwidth (between 20 days and 100 days) is much wider than that in the first mode. The 66-day peak is suggestive of activities related to the Caribbean anticyclones and LCFEs over the Campeche Bank continental slope. To have a better idea of the characteristics of oscillations, Fig. 13a shows a time–latitude plot of temperature distribution at 104.5 m along section C reconstructed from the second CEOF mode. Comparing with the plot constructed from the numerical simulation result (Fig. 13b, which is based on time series in Fig. 10a processed using the same procedures as in CEOF analysis, i.e., high-pass filtered with a half-amplitude period of 100 days and normalized by their own standard deviations at every spatial point), one can see that the mode recovers important information from the latter. For example, northward propagation of cold features on the slope of the Campeche Bank can be clearly seen and their timings are related to the time of anticyclone passage through the YC (comparing the long LCFE tracks indicated by dashed–dotted lines in Fig. 13a and Fig. 13b, and also comparing them with that in Fig. 10a). Although there are differences between Figs. 13a and 13b, since the mode only explains a fraction of the variance, the similarity between basic structures indicates that the second CEOF mode of the temperature field is dominated by propagation of LCFEs around the edge of the LC.

Fig. 13.
Fig. 13.

Time–latitude plots of (a) temperature (°C) reconstructed from the second CEOF mode and (b) high-pass-filtered (filter half-amplitude period 100 days) temperature (°C) anomalies from Fig. 10a along the Campeche Bank continental slope (section C in Fig. 1) at 104.5-m depth for the 5-yr output period of the ⅙° Atlantic Ocean general ocean model. Also indicated are events of Caribbean anticyclonic eddy approaching the YC (marked by arrows). The two dashed–dotted lines indicate LCFE propagation along this section.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

5. Discussion

Cyclonic frontal eddies have been frequently observed to propagate along the outer edge of the LC in satellite infrared imagery (Vukovich and Maul 1985; Fratantoni et al. 1998; Walker et al. 2003, 2009). However, most of them were detected, at the earliest, along the LC northwestern boundary where the water depth is greater than 1000 m. LCFEs are difficult to detect near the YC and over the Campeche Bank continental slope because of their small size and weak thermal contrast in SST imagery (Walker et al. 2003). In addition, continuous monitoring of their movements may be obscured by cloud cover and SST seasonal change in the region. For similar reasons, there are no reports of Caribbean anticyclones in the existing literature using satellite infrared observation.

Walker et al. (2009) is one of few studies that, by careful analysis of SST and satellite altimetry SSH data, reported the detection of small and large LCFEs as far south as 22°N in water depths of 200 to 1000 m. In Fig. 14, we provide a new piece of evidence from the near-real-time global gridded SSH maps (Leben et al. 2002) that supports our LCFE generation hypothesis. On 15 February 2002 (Fig. 14a), a large Caribbean anticyclone (positive SSH) was located near the southern border of the YC while there was clear evidence of a weak cyclonic eddy (negative SSH) at the north side of the LC in the channel. After twelve days (Fig. 14b), part of the anticyclonic eddy had moved through the YC and the cyclonic eddy had moved downstream and intensified in magnitude. The observed northward motion of the anticyclonic/cyclonic pair through the YC is very similar to the numerical model result illustrated in Fig. 2. This event was also observed in Archiving, Validation, and Interpretation of Satellite Oceanographic data (AVISO)/Collecte Localisation Satellites (CLS) delayed-time, merged absolute dynamic topography (not shown) estimated from sea level anomalies processed by Segment Sol multimissions d'ALTimétrie, d'Orbitographie et de localisation précise (SSALTO)/Data Unification and Altimeter Combination System (DUACS) (Ducet et al. 2000; LeTraon et al. 2003) and added to the Rio et al. (2011) mean dynamic topography.

Fig. 14.
Fig. 14.

Near-real-time SSH (cm) pattern in the GOM and the Caribbean Sea on (a) 15 Feb 2002 showing an anticyclone near the YC and a weak cyclone along the western margin of the LC just north of the YC, and (b) 27 Feb 2002 showing the anticyclone within the YC and an intensified cyclone north of the anticyclone.

Citation: Journal of Physical Oceanography 43, 9; 10.1175/JPO-D-12-0227.1

It should be pointed out that in this ⅙° Atlantic Ocean general circulation model anticyclones passing through the YC can sometimes cause a sizable perturbation on the inshore side of the Yucatan Current, which may then evolve into a cyclonic frontal eddy and propagate downstream along the edge of the LC. However, not every LCFE was generated in this way. Many times an LCFE was found when no Caribbean anticyclone was seen to pass through the YC. These cyclonic eddies may be attributed to barotropic, baroclinic, or mixing instability, as suggested by various authors (Hurlburt and Thompson 1982; Hurlburt 1986; Chérubin et al. 2006).

Because of the causal relationship between the generation of the LCFE near the YC and the passage of Caribbean anticyclones, we speculate that the LCFEs in the numerical model may be caused by the mechanism of vortex spinup attributed to shear layer roll up (Rott 1956; Pullin 1978; Pullin and Perry 1980). This mechanism is often evoked to explain transient eddy generation in barotropic tidal jets near headlands or tidal inlets (van Senden and Imberger 1990; Signell and Geyer 1991; Wells and van Heijst 2003). When water flows out of a channel, the no-slip wall condition leads to the growth of a viscous boundary layer that contains large-amplitude vorticity. At the sharp corner, which is the northeast tip of the Yucatan Peninsula in our case, the flow separates and the detached sheet of strong vorticity rolls up on itself to create a cyclonic vortex at the left-hand side of the jet (i.e., the LC). The eddy, thus generated, may then propagate with a speed that depends on the vorticity interaction between itself and the jet (Stern and Flierl 1987). Further in-depth analysis and more numerical experiments are needed in order to confirm or reject this postulation.

The Caribbean eddies seen in this Atlantic Ocean general circulation model were linked with eddies formed outside the Caribbean at the confluence of the North Brazil Current and North Equatorial Countercurrent systems (Carton and Chao 1999). Both cyclonic and anticyclonic eddies appear in the numerical model. Once generated, their fate as they propagate westward across the Caribbean and into the Yucatan Channel is governed by model dynamics. We find that sizable eddies (radius 100 km and larger) that survive to the western Caribbean and YC are predominantly anticyclones. Matsuura and Yamagata (1982) showed that cyclones dispersed more rapidly than anticyclones. A cyclone of diameter O(300 km) disperses on a time scale of O(100 days). Thus, cyclones in this model do not survive the approximately 1-yr period to traverse across the Caribbean from east to west. Therefore, no cyclones were observed to squeeze through the YC.

6. Conclusions

Results from a primitive equation ⅙° and 37-level Atlantic Ocean general circulation numerical model are analyzed and show that one type of LCFE results from strong interactions between the Loop Current proper and northward-moving Caribbean anticyclones as they pass through the Yucatan Channel. The cyclonic eddies first appear over the continental slope near the YC. Their typical swirl velocities are on the order of 40 cm s−1 and their vertical structures are quite coherent from near the sea surface to ~1000 m. Along the Campeche Bank continental slope they can reach the ocean bottom, while in the deep Gulf of Mexico region the isotherm doming can be seen below 1000 m depth. The source water inside these LCFEs can be traced back, using a backward-in-time Lagrangian particle-tracking method, to the western edge of the Caribbean Current in the Caribbean Sea and to the coastal water near the northern Yucatan Peninsula. More in situ hydrographic and float observations are needed to confirm or reject the existence, movement, and subsurface structure of the LCFE near the YC and over the Campeche Bank continental slope.

Time series analysis further reveals that anticyclonic eddy passage events, with a period of ~70 days, are dominant fluctuations in the EKE of the normal velocity component across the YC. These events seem to have no obvious effect on the large-scale LC dynamics and the timing of warm-ring shedding. However, they exert an influence on the net volume transport in the YC and on the variability of the volume-averaged relative vorticity in the LC region. The good correspondence between the timing of anticyclone passage and the peaks in the high-pass-filtered cyclonic relative vorticity series suggests that, accompanying the impingement of the Caribbean eddy, the maximum velocity of the LC and, subsequently, the cyclonic vorticity of the current will increase coincidentally. The linkage between the LCFE and the Caribbean anticyclone is first shown in a case study of the life cycle of a typical LCFE. It is then further illustrated in the time–latitude plots of temperature and relative vorticity along section C at 104.5 m. As shown, some LCFE events can be traced to Caribbean anticyclonic eddy events in which the generation of an LCFE occurs immediately after the Caribbean eddy enters the GOM through the YC. The reverse is not true. Sometimes when a Caribbean anticyclone enters the GOM, no LCFE is found or the LCFE induced is so weak that it quickly dissipates on the Yucatan Slope. The afore-mentioned phenomena is prominent in the ⅙° numerical simulation, as CEOF analysis of the temperature distribution at 104.5 m shows that the largest variance in the high-frequency band (i.e., period short than 100 days) is dominated by the Caribbean anticyclone events (i.e., the first CEOF mode) and the LCFE movement along the edge of the LC (i.e., the second CEOF mode). Near-real-time satellite altimetry imagery provides one piece of evidence that supports the proposed LCFE generation hypothesis. More systematical analyses of long time series of altimeter SSH are needed to further verify the hypothesis and reveal the frequency and possible causal relationship of the anticyclone/cyclone pair.

Acknowledgments

The research was supported by the Minerals Management Service Cooperative Agreement 14-35-0001-30804 and by the Office of Naval Research Grant N00014-00-1-0406. The numerical modeling research described in this publication was carried out, in part, at the Jet Propulsion Laboratory (JPL), California Institute of Technology, under a contract with the National Aeronautics and Space Administration (NASA). Computations were performed on the supercomputer provided through the JPL Supercomputing Project. We thank Chet Pilley at Earth Scan Laboratory, Louisiana State University, for processing the altimetry sea surface height images.

REFERENCES

  • Abascal, A. J., , J. Sheinbaum, , J. Candela, , J. Ochoa, , and A. Badan, 2003: Analysis of flow variability in the Yucatan Channel. J. Geophys. Res., 108, 3381, doi:10.1029/2003JC001922.

    • Search Google Scholar
    • Export Citation
  • Athie, G., , J. Candela, , J. Ochoa, , and J. Sheinbaum, 2012: Impact of Caribbean cyclones on the detachment of Loop Current anticyclones. J. Geophys. Res., 117, C03018, doi:10.1029/2011JC007090.

    • Search Google Scholar
    • Export Citation
  • Auer, S. J., 1987: Five-year climatological survey of the Gulf Stream system and its associated rings. J. Geophys. Res., 92 (C11), 11 70911 726.

    • Search Google Scholar
    • Export Citation
  • Barnett, T. P., 1983: Interaction of the monsoon and Pacific trade wind system at interannual time scale: Part I. Mon. Wea. Rev., 111, 756773.

    • Search Google Scholar
    • Export Citation
  • Barnier, B., , L. Siefridt, , and P. Marchesiello, 1995: Thermal forcing for a global ocean circulation model from a three-year climatology of ECMWF analysis. J. Mar. Syst., 6, 363380.

    • Search Google Scholar
    • Export Citation
  • Batchelder, H., 2006: Forward-in-time-/backward-in-time-trajectory (FITT/BITT) modeling of particles and organisms in the coastal ocean. J. Atmos. Oceanic Technol., 23, 727741.

    • Search Google Scholar
    • Export Citation
  • Candela, J., , J. Sheinbaum, , J. Ochoa, , A. Badan, , and R. Leben, 2002: The potential vorticity flux through the Yucatan Channel and the Loop Current in the Gulf of Mexico. Geophys. Res. Lett., 29, 2059, doi:10.1029/2002GL015587.

    • Search Google Scholar
    • Export Citation
  • Candela, J., , S. Tanahara, , M. Crepon, , B. Barnier, , and J. Sheinbaum, 2003: Yucatan Channel flow: Observations versus CLIPPER ATL6 and MERCATOR PAM models. J. Geophys. Res., 108, 3385, doi:10.1029/2003JC001961.

    • Search Google Scholar
    • Export Citation
  • Carton, J. A., , and Y. Chao, 1999: Caribbean Sea eddies inferred from TOPEX/Poseidon altimetry and a ⅙° Atlantic Ocean model simulation. J. Geophys. Res., 104 (C4), 77437752.

    • Search Google Scholar
    • Export Citation
  • Chao, Y., , and M. S. Lozier, 2001: Evaluation of North Atlantic property field simulations at ⅙°. J. Phys. Oceanogr., 31, 32003213.

    • Search Google Scholar
    • Export Citation
  • Chao, Y., , A. Gangopadhyay, , F. Bryan, , and W. R. Holland, 1996: Modeling the Gulf Stream system: How far from reality? Geophys. Res. Lett., 23, 31553158.

    • Search Google Scholar
    • Export Citation
  • Chérubin, L. M., , Y. Morel, , and E. P. Chassignet, 2006: Loop Current ring shedding: The formation of cyclones and the effect of topography. J. Phys. Oceanogr., 36, 569591.

    • Search Google Scholar
    • Export Citation
  • Cochrane, J. D., 1972: Separation of an anticyclone and subsequent developments in the Loop Current. Contributions on the Physical Oceanography of the Gulf of Mexico, L. R. A. Capuro and J. L. Reid, Eds., Gulf Publishing, 91–106.

    • Search Google Scholar
    • Export Citation
  • Ducet, N., , P. Y. LeTraon, , and G. Reverdin, 2000: Global high-resolution mapping of ocean circulation from TOPEX/Poseidon and ERS-1 and -2. J. Geophys. Res., 105, 477498.

    • Search Google Scholar
    • Export Citation
  • Dukowicz, J. K., , and R. D. Smith, 1994: Implicit free-surface method for the Bryan–Cox–Semtner ocean model. J. Geophys. Res., 99 (C4), 79918014.

    • Search Google Scholar
    • Export Citation
  • Elliott, B. A., 1982: Anticyclonic rings in the Gulf of Mexico. J. Phys. Oceanogr., 12, 12921309.

  • 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), 24 75924 779.

    • Search Google Scholar
    • Export Citation
  • Hofmann, E. E., , K. S. Hedstrom, , J. R. Moisan, , D. B. Haidvogel, , and D. L. Mackas, 1991: Use of simulated drifter tracks to investigate general transport patterns and residence times in the coastal transition zone. J. Geophys. Res., 96 (C8), 15 04115 052.

    • Search Google Scholar
    • Export Citation
  • Horel, J. D., 1984: Complex principal component analysis: Theory and examples. J. Climate Appl. Meteor., 23, 16601673.

  • Hurlburt, H. E., 1986: Dynamic transfer of simulated altimeter data into subsurface information by a numerical ocean model. J. Geophys. Res., 91 (C2), 23722400.

    • Search Google Scholar
    • Export Citation
  • Hurlburt, H. E., , and J. D. Thompson, 1982: The dynamics of Loop Current and shed eddies in a numerical model of the Gulf of Mexico. Hydrodynamics of Semi-Enclosed Seas, J. C. J. Nihoul, Ed., Elsevier Scientific, 243–297.

    • Search Google Scholar
    • Export Citation
  • Leben, R., , G. Born, , and B. Engebrecht, 2002: Operational altimeter data processing for mesoscale monitoring. Mar. Geod., 25, 318.

  • Leipper, D. F., 1970: A sequence of current patterns in the Gulf of Mexico. J. Geophys. Res., 75, 637657.

  • LeTraon, P. Y., , Y. Faugère, , F. Hernandez, , F. M. J. Dorandeu, , and M. Ablain, 2003: Can we merge Geosat follow-on with TOPEX/Poseidon and ERS-2 for an improved description of the ocean circulation? J. Atmos. Oceanic Technol., 20, 889895.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., , and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 117 pp.

  • Levitus, S., , R. Burgett, , and T. P. Boyer, 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 99 pp.

  • Lugo-Fernandez, A., 2007: Is the Loop Current a chaotic oscillator? J. Phys. Oceanogr., 37, 14551469.

  • Matsuura, T., , and T. Yamagata, 1982: On the evolution of nonlinear planetary eddies larger than the radius of deformation. J. Phys. Oceanogr., 12, 440456.

    • Search Google Scholar
    • Export Citation
  • Maul, G. A., 1977: The annual cycle of the Gulf Loop Current. Part I: Observations during a one-year time series. J. Mar. Res., 35, 2947.

    • Search Google Scholar
    • Export Citation
  • Maul, G. A., , and F. M. Vukovich, 1993: The relationship between variations in the Gulf of Mexico Loop Current and Straits of Florida volume transport. J. Phys. Oceanogr., 23, 785796.

    • Search Google Scholar
    • Export Citation
  • Merino, M., 1997: Upwelling on the Yucatan Shelf: Hydrographic evidence. J. Mar. Syst., 13, 101121.

  • Molinari, R. L., , J. F. Festa, , and D. W. Behringer, 1978: The circulation in the Gulf of Mexico derived from estimated dynamic height fields. J. Phys. Oceanogr., 8, 987996.

    • Search Google Scholar
    • Export Citation
  • Murphy, S. J., , H. E. Hurlbert, , and J. J. O'Brien, 1999: The connectivity of eddy variability in the Caribbean Sea, the Gulf of Mexico, and the Atlantic Ocean. J. Geophys. Res., 104 (C1), 14311453.

    • Search Google Scholar
    • Export Citation
  • Nakamura, M., , and Y. Chao, 2001: Diagnoses of an eddy-resolving Atlantic Ocean model simulation in the vicinity of the Gulf Stream. Part I: Potential vorticity. J. Phys. Oceanogr., 31, 353378.

    • Search Google Scholar
    • Export Citation
  • Ochoa, J., , J. Sheinbaum, , A. Baden, , J. Candela, , and D. Wilson, 2001: Geostrophy via potential vorticity inversion in the Yucatan Channel. J. Mar. Res., 59, 725747.

    • Search Google Scholar
    • Export Citation
  • Oey, L.-Y., , H.-C. Lee, , and W. J. Schmitz Jr., 2003: Effects of winds and Caribbean eddies on the frequency of Loop Current eddy shedding: A numerical model study. J. Geophys. Res., 108, 3324, doi:10.1029/2002JC001698.

    • Search Google Scholar
    • Export Citation
  • 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, 96399652.

    • Search Google Scholar
    • Export Citation
  • Pichevin, T., , and D. Nof, 1997: The momentum imbalance paradox. Tellus, 49, 298319.

  • Pullin, D. I., 1978: The large-scale structure of unsteady self-similar rolled-up vortex sheets. J. Fluid Mech., 88, 401430.

  • Pullin, D. I., , and A. E. Perry, 1980: Some flow visualization experiments on the starting vortex. J. Fluid Mech., 97, 239255.

  • Rio, M. H., , S. Guinehut, , and G. Larnicol, 2011: New CNES-CLS09 global mean dynamic topography computed from the combination of GRACE data, altimetry, and in situ measurements. J. Geophys. Res., 116, C07018, doi:10.1029/2010JC006505.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, L. K., 1985: Introduction to the CODE-2 moored array and large-scale data report. CODE-2 Moored Array and Large-Scale Data Rep., R. Limeburner, Ed., CODE Tech. Rep. 38, WHOI Tech. Rep. 85–35, 21 pp.

  • Rott, N., 1956: Diffraction of a weak shock with vortex generation. J. Fluid Mech., 1, 111128.

  • Schmitz, W. J., Jr., , and M. S. McCartney, 1993: On the North Atlantic circulation. Rev. Geophys., 31, 2949.

  • Sheinbaum, J., , J. Candela, , A. Baden, , and J. Ochoa, 2002: Flow structure and transports in the Yucatan Channel. Geophys. Res. Lett., 29, 1040, doi:10.1029/2001GL013990.

    • Search Google Scholar
    • Export Citation
  • Signell, R. P., , and W. R. Geyer, 1991: Transient eddy formation around headlands. J. Geophys. Res., 96 (C2), 25612575.

  • Stern, M. E., , and G. R. Flierl, 1987: On the interaction of a vortex with a shear flow. J. Geophys. Res., 92 (C10), 10 73310 744.

  • Sturges, W., 1992: An estimate of the spectrum of Loop Current variability from gappy data. J. Phys. Oceanogr., 22, 12451256.

  • Sturges, W., , and J. C. Evans, 1983: On the variability of the Loop Current in the Gulf of Mexico. J. Mar. Res., 41, 639653.

  • Sturges, W., , and R. Leben, 2000: Frequency of ring separation from the Loop Current in the Gulf of Mexico: A revised estimate. J. Phys. Oceanogr., 30, 18141819.

    • Search Google Scholar
    • Export Citation
  • Sturges, W., , N. G. Hoffmann, , and R. R. Leben, 2010: A trigger mechanism for Loop Current ring separations. J. Phys. Oceanogr., 40, 900913.

    • Search Google Scholar
    • Export Citation
  • van Senden, D. C., , and J. Imberger, 1990: Effects of initial conditions and Ekman suction on tidal outflows from inlets. J. Geophys. Res., 95 (C8), 13 37313 391.

    • Search Google Scholar
    • Export Citation
  • Vukovich, F. M., 1988: On the formation of elongated cold perturbation off the Dry Tortugas. J. Phys. Oceanogr., 18, 10511059.

  • Vukovich, F. M., 1995: An updated evaluation of the Loop Current eddy-shedding frequency. J. Geophys. Res., 100 (C5), 86558659.

  • Vukovich, F. M., , and G. A. Maul, 1985: Cyclonic eddies in the eastern Gulf of Mexico. J. Phys. Oceanogr., 15, 105117.

  • Walker, N. D., , S. Myint, , A. Babin, , and A. Haag, 2003: Advances in satellite radiometry for the surveillance of surface temperatures, ocean eddies and upwelling processes in the Gulf of Mexico using GOES-8 measurements during summer. Geophys. Res. Lett., 30, 1854, doi:10.1029/2003GL017555.

    • Search Google Scholar
    • Export Citation
  • Walker, N. D., , R. R. Leben, , S. Anderson, , J. Feeney, , P. Coholan, , and N. Sharma, 2009: Study of Loop Current frontal eddies based on satellite remote sensing and drifter data. U.S. Department of the Interior, Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA, OCS Study MMS 2009-023, 78 pp.

  • Walker, N. D., and Coauthors, 2011: Impacts of a Loop Current frontal eddy cyclone and wind forcing on the 2010 Gulf of Mexico oil spill. Monitoring and Modeling the Deepwater Horizon Oil Spill: A Record-Breaking Enterprise, Geophys. Monogr., Vol. 195, Amer. Geophys. Union, 103–116.

  • Wallace, J. M., , and R. E. Dickinson, 1972: Empirical orthogonal representation of time series in the frequency domain: I. Theoretical considerations. J. Appl. Meteor., 11, 887892.

    • Search Google Scholar
    • Export Citation
  • Wells, M. G., , and G.-J. F. van Heijst, 2003: A model of tidal flushing of an estuary by dipole formation. Dyn. Atmos. Oceans, 37, 223244.

    • Search Google Scholar
    • Export Citation
  • Zavala-Hidalgo, J., , S. L. Morey, , and J. J. O'Brien, 2003: Cyclonic eddies northeast of the Campeche Bank from altimetry data. J. Phys. Oceanogr., 33, 623629.

    • Search Google Scholar
    • Export Citation
Save