Clusters Of Drifters Research Articles

Near-surface horizontal convergence and dispersion near the tidal-mixing front on Northeastern Georges Bank

Horizontal dispersion and convergence are evaluated from near-surface drifter data collected during the summer and autumn of 1988 and summer of 1989 in the vicinity of the tidal-mixing front on northeastern Georges Bank. The effective horizontal diffusivity coefficient, K e, estimated from the relative motion of the drifters over the duration of the cluster tracking, increases as the cluster length scale, with a range from approximately −100 to 450 m 2 s −1 for length scales of approximately 5 to 58 km. At length scales<20 km, the magnitude of K e is near zero with many negative estimates, indicative of convergence. At length scales >20 km, K e increases in magnitude and scatter. Above 30 km length scales, K e is typically between 200–400 m 2 s −1. The dispersion as a function of length scale is compared to published estimates using apparent diffusivities, which are adjusted for the finite size of the drifter cluster at release. Georges Bank shows lower apparent diffusivities at length scales<20 km and higher ones at scales >20 km. The latter are believed to be principally due to the large horizontal gradients in the seasonal flow field, although other processes perhaps associated with the large tidal currents may also contribute. At the shorter length scales, the low dispersion is due to convergent processes. Examples of convergence at different locations and under different wind and stratification conditions are presented. Convergence near the tidal-mixing front tends to occur under light wind conditions and was observed in both seasons and in both years. During two intense southwestward windstorms, drifters also converged in the Bank's vertically well-mixed zone, apparently due to a combination of spatially varying Ekman depth and topographic effects.

The Shear, Convergence, and Thermohaline Structure of a Front*

With the objective of measuring convergence directly to confirm previous observations of frontal subduction, a seaward upwelling jet off central California was studied using satellite infrared images, hydrographic sections, ship drift, and clusters of surface drifters. The cyclonic front of the jet was sharper than 1 km, resulting in a shear several times larger than f. A cross-frontal convergence of 7 cm s−1 over the width of the front (equivalent to 0.8f) was visible as a 20-m-wide accumulation of debris. The sharpness of the front lasted at least for a day. Away from the cyclonic front, the divergence of the flow was small and the shear was less than 0.6f. Thermohaline layers, originating at the front, were interleaving along isopycnals, suggesting water subduction. It is proposed that the asymetry between anticyclonic and cyclonic sides of the jet, and the strong convergence at the cyclonicfront, resulted from a frictionally driven ageostrophic secondary circulation superimposed on the geostrophic flow.

Circulation in the Ensenada Front—September 1988

The 3-D circulation of the Ensenada Front, located off southern California and northern Baja California, is examined using hydrographic and mixed-layer drifter measurements. The measurements resolve sub-mesoscle motions (10 km scale) and trace the flow over a 400 km extent, from deep water into the continental margin. Along the Ensenada Front, the cool, relatively fresh, southeastward-flowing waters of the California Current were observed to turn abruptly onshore and flow eastward for approximately 200 km before resuming a southeastward trajectory. Embedded in the onshore flow was a 60 km wide, surface-intensified (> 30 cm s −1), high speed filament, which bifurcated after resuming a southeastward course, and appeared to feed into a complex set of flows that were strongly influenced by the mesoscale eddy field. Various water mass properties are used to trace the along-front evolution of the filament and its demise. Estimates of vorticity and divergence in the filament are made on the scale of drifter clusters, deployed in diamond patterns with a separation of 10 km. Potential vorticity appears to be conserved along drifter trajectories, and suggests a divergent, upwelling (5 m day −1) filament with convergent flow along the sides. Such a pattern is consistent with a model of constant wind stress blowing along the filament.

Measurements of drifter cluster dispersion

Abstract The motion of clusters of drifters in stratified coastal waters was studied. Wave‐like motion dominated the single‐particle statistics and had length scales larger than the cluster dimensions; consequently it was largely filtered out of motion relative to the cluster centroid. But the stochastic motion causing eddy diffusion seemed to be equally present in both single‐particle motion and motion relative to the cluster centroid. The single‐particle kinetic energy was 10 and 2 times the kinetic energy of motion relative to the centroid in summer and winter, respectively. The relative motion had longer Lagrangian integral time‐scales and smaller Eulerian spatial correlation scales than the single‐particle motion. Integral length scales of relative motion were of 0.1 and 0.2 times the standard deviation of drifter positions about the centroid for summer and winter ensembles, respectively. The y component of ensemble‐averaged relative eddy diffusivity in summer (0.21 m2 s−1) was much larger than that ...

Patch diffusion computed from Lagrangian data, with application to the atlantic equatorial undercurrent

Abstract Radar‐tracked drifter trajectories reported by Fahrbach et al. (1986) are reanalysed. The data are obtained from 4 drifter clusters that were deployed in the Atlantic Equatorial Undercurrent for approximately 48 h each. The ratio of Lagrangian integral length scale to the standard deviation of drifter positions ranged from C = 0.02 to 0.36. However, on average it was 0.12 in the meridional direction, thus confirming Fahrbach et al.’s (1986) calculations of the meridional diffusion of salt from the Atlantic Equatorial Undercurrent. Making a quasi‐stationary assumption we find that relative velocities had Lagrangian integral time‐scales of about 1 h when relative velocities were large O(0.1 m s−1) and about 3 h when they were smaller O(0.05 ms−1 ). The Eulerian integral time‐scale for relative velocities is only slightly greater than the Lagrangian integral time‐scale. Joint space‐time correlations indicated that the eddy velocity changes rapidly compared with time‐scales required to advect a particle over the eddy length scale. In the case of D1, D2 and D3 experiments (which had large relative velocities) the relative velocity became spatially decorrelated at scales small compared with cluster dimensions. The spatial scales of the relative velocity were somewhat larger for the CIPREA experiment, which had weaker relative velocities than D1, D2 and D3.

The fractal dimension of relative lagrangian motion

14 clusters of drifters are studied to obtain the fractal dimension of relative trajectories between pairs of drifters. Averaging over all experiments, the relative trajectories have a fractal dimension of 1.3 over separation scales ranging from 10 m to 4000 m. The clusters were deployed in Lake Erie, the Atlantic Equatorial Undercurrent, and in coastal waters off the south shore of Long Island. The average fractal dimension was quite similar for groups of experiments in each of these 3 areas, but it fluctuates over a greater range (1.59 to 1.12) when we consider individual experiments. This and consideration of variation of the generalized fractal dimension for higher moments, raise the possibility that trajectories are multifractal. The estimates we obtain for fractal dimension of relative lagrangian motion should be regarded as first estimates, pending analysis of more substantial data sets. DOI: 10.1034/j.1600-0870.1990.t01-4-00005.x

A model for the analysis of drifter data with an application to a warm core ring in the Gulf of Mexico

A model developed primarily for the analysis of drifter data is discussed. The model is based on a parametric representation of the drifter velocity in terms of Monge potentials, two of which are constants following the motion of the fluid. In the application presented here, one of the potentials is taken as the frequency of rotation about a ring by a parcel. A more restricted interpretation of the velocity gradient in variant for horizontal flow is also discussed. The other potential is taken as the streamline of a parcel that is locally approximated as a conic section. For the restricted interpretation, solutions to the model equations are critically dependent on the relative sizes of the squares of the vertical vorticity and the total deformation rate. This model differs from drifter cluster models in that each drifter provides independent estimates of the vorticity and deformation rates. Application is made to path data from three drifters that were seeded in a warm core ring in the Gulf of Mexico in November 1980. The model provided estimates of the ring translation and swirl velocities along with the ring geometry. The analysis showed that the ring was persistently elliptical with the major axis aligned in the east‐west direction. A satellite infrared photo on January 21,1981, confirmed this orientation.

Effect of Sampling Rate and Random Position Error on Analysis of Drifter Data

The central question discussed here is how the rate at which drifter positions are determined and the position errors affect the calculation of velocity, acceleration and velocity gradients such as divergence and vorticity. The analysis shows that the mean-square velocity and acceleration errors each are composed of two terms. One arises from the position uncertainty and the discrete sampling rate. The other term is an alias resulting from sampling a continuous velocity or acceleration spectrum discretely. Effects at low and high frequencies and sampling intervals are examined by asymptotic expansions of the spectra. Then optimum smoothing and derivative filters are obtained for the velocity and accelerations, respectively. The efficiency of these filters is determined by comparison with the errors previously established. The calculation of divergence and vorticity from drifter clusters typically neglects the position error, in which case the errors in the velocity gradients are proportional to the velocity errors. Our analysis shows that this procedure produces estimates of the velocity gradients whose magnitudes are less than the true values. This bias is easily removed. The analysis is concluded with a derivation of formulas for unbiased estimates of the variance and covariance of the velocity gradients.

Calculations of Differential Kinematic Properties from Lagrangian Observations in the Western Caribbean Sea

Observations of the motions of drifter clusters were made in the western Caribbean Sea during the summer of 1971. By two independent analyses of the relative motions of a cluster, two time series of horizontal divergence, vorticity, shear deformation rate, and normal deformation rate are developed. The results of the two approaches are very similar. The time series for these differential kinematic properties are fairly smooth when the drifters were moving in the Yucatan Current. Otherwise, the time series are ragged with frequent changes in sign. It is speculated that the raggedness is due to small values of the shear rates relative to random observational errors or small-scale turbulent processes. The records of the differential kinematic properties are used to evaluate the stretching and material derivative terms of the vorticity equation. Calculations indicate that potential vorticity is conserved along trajectories in the Yucatan Current.