BIBLIOGRAPHY
3.1. RADIO BUOYS
3.1.4. GPS-Tracked Drifters
Figure 3.9shows a plot of the trajectory of a satellite- tracked drifting buoy deployed off Goa, India.
Starting on 15 March 2011, the Argos user communities were provided the option to choose between two location processing algorithms:
l The algorithm based on the classical least-squares method that has been employed since Argos process- ing began in 1986
l The algorithm based on Kalman filtering
The trajectory of an Argo float in the Pacific calculated with the algorithm based on Kalman filtering and least- squares analysis is given inFigure 3.10.
of 1 min) scales of interest in surf zone circulation are re- solved with GPS techniques used in coastal and near-ship drifters (e.g., real-time DGPS, 3e5-m accuracy [George and Largier, 1996] and carrier phase post-processing, 1-m accuracy [Doutt et al., 1998]). Drifters that autonomously record their position enable high-frequency Lagrangian data to be collected.
Near-shore drifters are intended for use in confined or near-shore environments over timescales of up to several days and are a low-cost alternative for applications that do not require drifters with full ocean-going capabilities.
According to George and Largier (1996), near-shore drifters must be inexpensive so that large numbers can be deployed simultaneously to provide both statistical reli- ability and spatial resolution over the domain of interest.
The system must be self-contained, reliable, rugged, and easily transportable. The drifters are to be retrievable. The drifter package should be modular, allowing for the possi- bility of using different drogue, float, and visibility device designs. The drifter should be easily adjusted for drogue at depths from 1 to 15 m for use in vertically sheared flow in a wide variety of applications. Positioning error must be small (less than 5 m) in order to resolve the small-scale structure of the flows of interest. Sample interval must be short (less than 10 m) and not dependent on the number of drifters deployed at one time. Position accuracy and sample rate must be maintained at distances of up to 20 km from the base of operations. In the presence of moderate wind and wave conditions, the errors in the water-following performance of the drifter must be known and the wind- induced drift-to-wind-speed ratio must not exceed 0.005.
A near-shore GPS drifter constructed of canvas moun- ted on a fiberglass frame has been described byGeorge and Largier (1996). The projected area of this drogue is approximately 1 m2, and the electronics package is carried
in a central PVC pressure case. This drifter has known surface-following characteristics in a wave field, has low mass, and can be inexpensively constructed. However, high impacts from breaking waves and the tendency to “surf”
shoreward often precluded the use of this device.
Schmidt et al. (2003)described a surf zone drifter and results from a deployment near a rip current. In this system, an impact-resistant body of tubular PVC is ballasted for nearly complete submergence (Figure 3.11). They cir- cumvented the problem of surfing by attaching a PVC disk to the base of their 50-cm-long main casing. This disc strongly dampens the vertical response of the drifter, allowing broken and near-breaking waves to pass over without rapidly pushing the drifter ashore. This mechanism was found to be effective in resisting surfing. Their longer receiver casing also penetrates deeper into the water below, breaking wave rollers, and thereby precludes some of the need for an additional drogue. The heave and roll of the drifter, which can degrade GPS position estimates by causing large, rapid deviations of the antenna orientation from vertical, are reduced by the damping plate and by the vertical separation of the centers of buoyancy and mass.
A shoreward mass flux is associated with breaking waves and bores, but by design, bores and breaking waves pass over the drifters.
In the design of Schmidt et al. (2003), each drifter records GPS pseudo-range and carrier phase data at 1-Hz for post-processing and transmits this information to shore at 10-s intervals for real-time differential GPS (DGPS) tracking and partial data backup (Figure 3.12). The telemetry range is about 5 km for a shore antenna elevation of 10 m. Battery storage limits the deployment duration to 24 h. Figure 3.13 shows a conglomeration of drifter trajectories observed in a surf zone. Field intercomparison measurements have indicated that the mean alongshore
FIGURE 3.10 Trajectory of an Argo float in the Pacific Ocean calculated with the algorithm based on Kalman filtering (top image) and least-squares analysis (bottom image).(Source:
ARGOS SYSTEM newsletter, FLASH #20, February 2011,ÓCLS 2012, reproduced with kind permission of CLS Service Argos, Toulouse Cedex, France.)
101 Chapter | 3 Lagrangian-Style Surface Current Measurements Through Tracking of Surface Drifters
currents estimated from trajectories of the 0.5-m-draft drifters in 1e2-m water depth and nearby fixed current meters within the surf zone agreed well with measurements obtained from nearby bottom-mounted acoustic current meters (correlation 0.95 and rms differences of 10 cm/s).
Drifters deployed near a rip current often followed eddy- like trajectories before being advected seaward of the surf zone (Figure 3.14).
InJohnson et al. (2003), a simple, robust receiver unit suitable for use in environments such as the near-shore zone, lakes, and estuaries was described, and specialized drogue arrangement for use in the surf zone was briefly discussed.
Johnson and Pattiaratchi (2004a) reported a simple, low-cost drifter design for surf zone applications based on nondifferential GPS position fixing and having the capa- bility to collect high-frequency, accurate, Lagrangian data.
They also addressed the main issues of the dynamic response of the drifters, a robust design capable of auton- omous position fixing, position-fixing accuracy, and field validation.
The solution found byJohnson and Pattiaratchi (2004a) to overcome the challenging environment of the surf zone was to use a series of “soft” drogues attached to a small, compact receiver unit. The drifter arrangement, shown in Figure 3.15, is a cylindrical receiver unit connected to a series of soft parachute drogue elements. This type of drogue opens and dramatically increases its drag when there is a differential velocity between the upper and lower part of the water column, as is the case in wave breaking.
The parachute drogue also stabilizes the drifter and
FIGURE 3.12 Drifter system schematic. The base station performs DGPS and data-logging functions and can serve a fleet of 10 drifters. A mobile tracking station monitors drifter DGPS positions in real time.(Source: Schmidt, W. E., B. T. Woodward, K. S. Millikan, R. T. Guza, B. Raubenheimer, and S. Elgar: A GPS- tracked surf zone drifter, J. Atmos. Ocean. Technol., 2003 (20) 1069e1075. ÓAmerican Meteoro- logical Society. Reprinted with permission.)
FIGURE 3.11 Schematic of a surf zone drifter. Surface-piercing antennae for receiving GPS signals and for radio-frequency (RF) communication with shore are molded permanently to the drifter top cap.(Source: Schmidt, W. E., B. T. Woodward, K. S. Millikan, R. T. Guza, B. Raubenheimer, and S. Elgar: A GPS- tracked surf zone drifter, J. Atmos. Ocean. Technol., 2003 (20) 1069e1075.ÓAmerican Meteorological Society. Reprinted with permission.)
prevents it from rolling excessively. The drifter floats with only 2 cm of the receiver casing above the water, so the effect of windage (i.e., stress exerted by the wind directly on the float) is expected to be very small.
The receiver units are 32 cm long and 10 cm in diameter and obtain and record a GPS position fix at 1-second interval. These units consist of an integrated GPS antenna/receiver wired to a data logger and a power source in a highly robust waterproof housing. Details of their construction and the internal components can be found in Johnson and Pattiaratchi (2004b). They are deployed and recovered manually and can easily be used for repeated
runs over a short period of time; their ease of deployment makes them effective for measuring rapidly developing transient features. To minimize windage and inertial effects, the drifter units are designed for nearly neutral buoyancy so that only the upper surface covering the internal GPS antenna projects above the water. In calm water, only 2 cm of the instrument projects above the surface; therefore, it is reasonable to assume that windage is negligible. A wire with a small ribbon is extended above the main unit to serve as a visual aid to enhance visibility.
The drogue is a series of parachute-shaped elements that hang below the GPS drifter casing. The parachutes are made of a cone of Dacron sailcloth with webbing attach- ments and are almost neutrally buoyant. A small weight is attached at the end of the parachutes to keep them hanging below the receiver unit. When the receiver unit is pulled shoreward by the breaking section, the parachutes open and anchor the drifter to the orbital velocities below the breaking region. In nonbreaking waves, the parachutes are closed and hang almost vertical and therefore present only their cross-sectional area. The drogue has also been found to be very effective in stabilizing the receiver unit by strongly damping the oscillatory motions that an “un- drogued” receiver experiences. In water depths less than the length of the complete drifter arrangement, the drogue may touch the seabed, which inevitably causes some measure- ment error. This means that the minimum operating depth is
FIGURE 3.13 A conglomeration of drifter trajectories observed in a surf zone. Thin curves and filled circles represent drifter trajectories and ADV current meter locations, respectively.(Source: Schmidt, W. E., B. T. Woodward, K. S. Millikan, R. T. Guza, B. Raubenheimer, and S. Elgar: A GPS- tracked surf zone drifter, J. Atmos. Ocean. Technol., 2003 (20) 1069e1075.ÓAmerican Meteorological Society. Reprinted with permission. )
FIGURE 3.14 Trajectories of drifters released (locations shown with large circles) near the base of a rip current. The small filled circles on each trajectory indicate 120-s intervals.(Source: Schmidt, W. E., B. T. Wood- ward, K. S. Millikan, R. T. Guza, B. Raubenheimer, and S. Elgar: A GPS- tracked surf zone drifter, J. Atmos. Ocean. Technol., 2003 (20) 1069e1075.
ÓAmerican Meteorological Society. Reprinted with permission.)
FIGURE 3.15 Drifter specifically designed for the surf zone, with dro- gue attached.(Source:Johnson and Pattiaratchi, 2004a.)
103 Chapter | 3 Lagrangian-Style Surface Current Measurements Through Tracking of Surface Drifters
the length of the receiver unit (32 cm), though the drogue still provides sufficient resistance to surfing in large waves.
Careful visual observation byJohnson and Pattiaratchi (2004a)of the drifter in breaking waves indicates that the parachutes are extremely effective in resisting surfing. When in the overturning section of a plunging breaker, the para- chutes prevent the receiver from accelerating up the face and forward with the plunging lip. In spilling breakers, the receiver unit ducks underneath the foaming roller section and reappears at the surface a couple of seconds later as the wave passes over. The only situation in which the drifters do not perform well is when they’re caught at the plunge point of a strongly plunging wave. However, due to the strong downward velocities, the whole assembly may be completely disrupted and rolled by the wave, the parachutes no longer correctly oriented, and the drifter then tends to surf in the developing roller section. When caught in strong breaking events, the drogue is ineffective at anchoring the drifter to the wave orbital motion and the whole unit travels at close to the wave phase speed. Surfing events are very easy to identify in the data, since the phase speed is much greater than orbital velocities of water particles and typical wave- averaged current speeds. Data contaminated by surfing can then easily be excluded from any analysis.
There are errors inherent in both the position fixing of the GPS receiver and in the calculation of velocities and accelerations from raw position data. Although the removal of SA has greatly improved the performance of non- differential GPS, there are still errors in the reported posi- tions due to various factors (Hofmann-Wellenhof et al., 1997). The magnitude of errors can be greatly reduced and decimeter accuracy obtained in a differential mode using either code phase or carrier phase data. However, this does introduce an additional level of complexity and significant additional cost into the drifter design.
?>In the surf zone, there are motions over a wide range of frequencies, and it is necessary to determine how much different frequency ranges are affected by the positioning errors. Error analysis byJohnson and Pattiaratchi (2004a) indicates that nondifferential GPS position fixing is suffi- cient for accurately measuring motions with frequencies below about 0.05 Hz. The relative RMS error can be greatly reduced using differential methods and effectively allows accurate positioning for motions at frequencies of 1 Hz (Schmidt et al., 2003).
Intercomparison studies (drifter versus vertical profile measurements using a bottom-mounted acoustic Doppler current profiler) have suggested that the drifters slightly underestimate the depth-averaged velocity in both direc- tions. It must, however, be borne in mind that direct vali- dation of the drifters in the surf zone is somewhat problematic. For example, whereas the drifter measures near-surface velocities, the profile instrument near the bed measures near-bed velocities. This highlights the difficulties
of comparing different types of surf zone measurements and the care required in assessing exactly what drifters are measuring. The type of drifter design, in terms of receiver and drogue arrangement, is also clearly important in deter- mining the response in the cross-shore direction. Unfortu- nately, in the absence of any “true” Lagrangian velocity information, it requires comparison of wave-averaged Lagrangian drifter velocities with fixed Eulerian depth- and wave-averaged data. Therefore, the two instruments are not actually measuring the same parameter. This is partic- ularly the case in the surf zone, because the speed and short length scales of typical surf zone currents mean that the two instruments are quickly separated and measure different parts of the water current field.
The cost, size, and weight of the GPS-based surf-zone drifter (based on the design of Schmidt et al., 2003) reported by MacMahan et al. (2009) allow for a large number of independent observations to be obtained for various meteorological and oceanographic applications. In a field experiment they conducted in May 2007 at Monterey Bay, California (a natural beach with persistent rip currents), a cluster of 10 drifters was simultaneously released with dye. Fluorescent dye represents a “true”
Lagrangian measurement of water movement. The bulk of the dye patch and the drifter cluster remained co-located for one rip-current circulation, suggesting that the drifters provide accurate Lagrangian estimates. Both the dye and the drifters were spreading slightly, but they remained together in a similar patch/cluster for the initial circulation.
The dye followed the average drifter velocity estimates and completed a revolution in 5 min, similar to the drifters. The experimental results enabled them to arrive at a confident conclusion that the drifters represent Lagrangian estimates.
The ability to obtain Lagrangian velocity observations in a rip-current system is a good example of the usefulness of the inexpensive handheld GPS. The flow field of rip currents has a large spatial variability that is difficult to measure with in situ instruments owing to expense and deployment complexities. The inexpensive system reported byMacMahan et al. (2009)has the ability to fill in the voids betweenin situinstruments, thereby advancing our under- standing of the hydrodynamics of a rip-current system.