George O. Marmorino

Empirical Orthogonal Function Analysis of Ocean Surface Currents Using Complex and Real-Vector Methods*

Empirical orthogonal function (EOF) analysis has been widely used in meteorology and oceanography to extract dominant modes of behavior in scalar and vector datasets. For analysis of two-dimensional vector fields, such as surface winds or currents, use of the complex EOF method has become widespread. In the present paper, this method is compared with a real-vector EOF method that apparently has previously been unused for current or wind fields in oceanography or meteorology. It is shown that these two methods differ primarily with respect to the concept of optimal representation. Further, the real-vector analysis can easily be extended to threedimensional vector fields, whereas the complex method cannot. To illustrate the differences between approaches, both methods are applied to Ocean Surface Current Radar data collected off Cape Hatteras, North Carolina, in June and July 1993. For this dataset, while the complex analysis ‘‘converges’’ in fewer modes, the real analysis is better able to isolate flows with wide cross-shelf structures such as tides.

Convergence and Downwelling at a River Plume Front

Abstract The small-scale structure of the circulation and hydrography at the frontal boundary of the Connecticut River plume in Long Island Sound has been resolved using a novel combination of instruments: a towed acoustic Doppler current profiler (ADCP) and a rigid array of current meters and conductivity–temperature sensors. Observations were made during the latter half of the eastward ebb tide, when the river plume was well established and the front was moving to the west at approximately 0.3 m s−1. Two across-front transects revealed a horizontal convergence rate in the across-front velocity components at 0.6 m of 0.05–0.1 s−1. This was associated with a salt-induced horizontal density gradient of 10−2 kg/m4. Observations obtained during a period in which the towed ADCP was caught in the zone of maximum surface convergence showed significant downwelling with a near-surface maximum of 0.2 m s−1. Vertical velocities of this magnitude are consistent with observed magnitudes of the convergence rate at 0.6...

An airborne, real aperture radar study of the Chesapeake Bay outflow plume

An airborne, real aperture radar (RAR) has been used to study the fronts associated with the Chesapeake Bay outflow plume during spring outflow conditions. The RAR produced images of the ocean surface with a range resolution of 10 m, an azimuthal resolution of approximately 30 m, and an image size of 2.5 km × 24 km. Two sampling strategies were utilized: one to synoptically map the entire mouth of the Chesapeake Bay at roughly hourly intervals; and a second to capture the rapid evolution of particular features. In addition, flight times were chosen such that over the course of the entire experiment, data were collected over all phases of the semidiurnal tidal cycle. Three distinct frontal signatures were observed in the imagery. A primary front extended from inside the estuary along the Chesapeake Channel to an anticyclonic turning region east of Cape Henry, and then extended southward along the coast toward Cape Hatteras. This is the classic expression of the plume front, inertial turning region, and coastal jet. A second front with a north-south orientation was observed approximately 20 km east of the bay mouth. This secondary front appears to mark the residual offshore density gradient. A third front was identified east and south of Cape Henry, within 2 km of the coast. This front appears to mark the inshore edge of the plume and has not been documented previously. Time sequences of the imagery indicate that when moving in a clockwise sense around the primary front, the frontal translation speed varies systematically from 20 cm/s in the northern section to 50 cm/s in the south. The position of the primary front and the locations and trajectories of small-scale frontal cusps suggest that bathymetry may be both a significant determinant of the front location as well as a source of along-front variability. These observations are possible due to the airborne RARs ability to collect high-frame rate image sequences, a capability that is not shared by present space-based radar systems.

An EOF analysis of HF Doppler radar current measurements of the Chesapeake Bay buoyant outflow

Surface currents measured by HF Doppler radar as part of a study of the Chesapeake Bay outflow plume are examined using a ‘real-vector’ empirical orthogonal function (EOF) analysis (Kaihatu et al., 1998). Based on about 23 days of nearly continuous data, the analysis shows that the first three EOF modes, judged to be the only significant modes, account for 76% of the variance in the data set. The buoyant outflow occurs primarily in the mean flow field. The first EOF mode is dominated by wind forcing and the second mode by across-shelf semi-diurnal tidal forcing. The third mode exhibits a large-scale horizontal shear and contains a curved region of weak relative flow which appears to delineate the o⁄shore edge of the plume; also, the third-mode response varies over the spring-neap cycle, suggesting a modulation of the outflow plume by a tidal residual eddy. The analysis therefore has provided a useful, exploratory examination of this dataset of surface currents. ( 1999 Elsevier Science Ltd. All rights reserved.

Gulf Stream surface convergence imaged by synthetic aperture radar

On July 20, 1990, the north edge of the Gulf Stream (36.7°N, 72.0°W) was sampled by the R/V Cape Henlopen and simultaneously imaged by the Jet Propulsion Laboratorys airborne synthetic aperture radar (SAR). Hydrographic measurements show an abrupt surface front separating warm, salty Gulf Stream water in the south from a filament of cool, fresh (<33 practical salinity unit (psu)) water to the north. The filament lies within the stream and is likely water entrained from the continental shelf. The southern boundary of the filament is marked by increased surface wave breaking in a 100- to 200-m-wide zone, accumulations of Sargassum, and an orthogonal velocity change of 20 cm/s. The front is manifested in a sequence of SAR images as a narrow line having returns 1–2 dB higher than background. (A second, transient SAR line occurs near the northern filament boundary.) The observations are compared with model calculations of the surface wave hydrodynamics and radar scattering. The ocean waves are driven by southwesterly 8-m/s winds and interact with the front to produce primarily an enhancement of 2- to 3-m waves over a ≲200-m-wide region centered downwind of the front. Using a composite scattering radar model along with measured breaking-wave statistics, we show that the observed modulations in the radar backscatter can be accounted for through breaking-wave and tilted Bragg wave scattering effects. These results further show that SAR images of the ocean surface can be exploited for detailed study of particular ocean processes.

Observations of the response of a buoyant estuarine plume to upwelling favorable winds

[1] The response of the buoyant plume from the Chesapeake Bay to upwelling favorable winds and during conditions of low freshwater discharge is examined using an offshore array of bottom-deployed acoustic Doppler current profilers (ADCPs) as well as ancillary data. During periods of northward (upwelling favorable) winds the observed response is strong offshore flow (U > 0.2 m/s) in a shallow layer of plume water and a rapid reversal of the near-surface flow from the usual southward flow associated with inertial turning of the plume. The offshore migration of the plume front is deduced from sequential arrivals at the various mooring locations. The orientation of the plume front is approximately alongshore (north-south), suggesting that the upwelling response is two-dimensional. A plume thickness calculated from the ADCP data is relatively constant at about 4 m as the plume moves across the moorings. The offshore propagation speed of the plume front of about 0.20 m/s matched the advection speed predicted by Ekman theory using the calculated plume layer thickness. The mooring velocity profiles provide a view of the across-plume structure in a reference frame fixed to the front. This reveals a pattern of circulatory motion with sinking of plume and ambient water at the front and an ascending motion occurring inshore, suggesting a recirculation between the seaward and inshore portions of the plume. These results are consistent with recent model simulations of the response of a river plume during upwelling favorable winds.

Calibrating a Gyrocompass Using ADCP and DGPS Data

Abstract Accurate values of ship’s heading, usually obtained from a gyrocompass, are vital in calculating absolute currents from ship-mounted acoustic Doppler current profilers (ADCPs). This note presents a straightforward method of calibrating a ship’s gyro by comparing two independent estimates of ship’s absolute velocity: one from bottom-track ADCP data (in combination with gyro data), and the other from differential global positioning satellite data. From a dataset collected in June 1993, 5-min-averaged data were isolated having these two estimates of ship’s velocity. These data showed that the speed estimates agreed within 0.2% but that the direction estimates varied by ±5° (standard deviation 2.4°) in a manner that was strongly a function of ship’s gyro heading. Correcting the data for this relationship reduced the standard deviation of the components of the ship-velocity differences by a factor of 2 and removed clear biases in their means.

Correlation of oceanographic signatures appearing in synthetic aperture radar and interferometric synthetic aperture radar imagery with in situ measurements

Synthetic aperture radar (SAR) imagery collected over the continental shelf near Cape Hatteras, N. C., is analyzed in conjunction with shipboard hydrographic and current measurements. The SAR measurements were made over a 2-hour period on June 20, 1993, in both standard mapping mode and interferometric synthetic aperture radar (INSAR) mode from a NASA DC-8 aircraft as part of the High-Resolution Remote Sensing Experiment. In situ currents were measured using a surface-towed acoustic Doppler current profiler (ADCP). The measurements were made near the end of a period of Gulf Stream incursion onto the shelf as detected using a shore-based HF radar. Winds were southwesterly at 4-6 m s -1 . Long, curvilinear SAR signatures, resembling earlier SEASAT observations made in the same area, are shown to correspond to narrow, shallow fronts separating water masses that increase in surface density with distance offshore. Across-front changes in surface current inferred from the INSAR data are consistent with 2-m-depth currents measured by the ADCP over scales of tens of meters. Thus frontal current gradients measured by INSAR reflect real changes in surface current and are not due to biases induced by changes in the surface-wave spectrum. This lends support to the detailed INSAR surface maps derived by Graber et al. [1996]. An east-west salinity front having the largest observed surface density and current gradient is corrugated on length scales up to the local Rossby radius of deformation and translates southward between successive images. In data from the longer radar wavelengths the salinity front appears as a dark band downwind of a bright signature, and this is interpreted as a region where Bragg-scale waves regenerate following their dissipation in the frontal region. In addition to the fronts the imagery shows closely spaced packets of southward propagating ocean internal waves occurring in the strongly stratified inshore water mass. This case study further serves to emphasize the potential of SAR imagery for study of a wide range of shelf processes.

Observations of an inshore front associated with the Chesapeake Bay outflow plume

Abstract Preliminary observations are reported of a recurring front located near Cape Henry, Virginia, USA. The front occurs on the right-hand side, looking seaward, of the buoyant plume discharging from the Chesapeake Bay and separates the plume from a band of relatively dense seawater confined against the Virginia coast. The front thus appears to be of a type similar to the inshore plume front reported by Sanders and Garvine for the Delaware Bay. Similar to an estuarine tidal intrusion front, the Cape Henry front evolves to a prominent V-shaped planform during flood tide and subduction of fluid along the front may provide a means for recirculation of near-surface material.

Measurement of surface currents using sequential synthetic aperture radar images of slick patterns near the edge of the Gulf Stream

Two-dimensional surface currents are estimated over an area of ∼100 km2 near the inshore edge of the Gulf Stream by correlating the surface slick patterns observed on two synthetic aperture radar (SAR) images collected about 20 min apart. The currents obtained from this analysis are found to agree well with shipboard acoustic Doppler current profiler (ADCP) measurements at 10 to 20-m depths. The orientations of the long, linear slicks observed in this region differ from the mean current direction by about 16° but coincide with the direction of maximum radial deformation or strain, as calculated from the observed current field. The calculated current gradients indicate a mean surface convergence of about 2×10−5 s−1 and a cyclonic vorticity of about 1×10−4 s−1, values which fall within the range of previous estimates in the study area. These results suggest that slick patterns may be used to infer surface current magnitude and direction from a time sequence of SAR images but that care must be exercised in inferring current directions from the orientations of slick patterns within a single image.