Search

Surface drifters and virtual particles are used to investigate transport between seven coastal regions in the central and southern Adriatic Sea to estimate the degree to which these regions function as a network. Alongshore coastal currents and cyclonic gyres are the primary circulation features that connected regions in the Adriatic Sea. The historical drifter observations span 25 years and, thus, provide estimates of transport between regions realized by the mean surface circulation. The virtual particle trajectories and a dedicated drifter experiment show that southeasterly Sirocco winds can drive eastward cross-Adriatic transport from the Italian coast near the Gargano Promontory to the Dalmatian Islands in Croatia. Southeasterly winds disrupt alongshore transport on the west coast. Northwesterly Mistral winds enhanced east-to-west transport and resulted in stronger southeastward coastal currents in the western Adriatic current (WAC) and export to the northern Ionian Sea. The central Italian regions showed strong connections from north to south, likely realized by alongshore transport in the WAC. Alongshore, downstream transport was weaker on the east coast, likely due to the more complex topography introduced by the Dalmatian Islands of Croatia. Cross-Adriatic connection percentages were higher for east-to-west transport. Cross-Adriatic transport, in general, occurred via the cyclonic sub-gyres, with westward (eastward) transport observed in the northern (southern) arms of the central and southern gyres. (C) 2016 Elsevier Ltd. All rights reserved.

Ocean prediction systems rely on an array of assumptions to optimize their data assimilation schemes. Many of these remain untested, especially at smaller scales, because sufficiently dense observations are very rare. A set of 295 drifters deployed in July 2012 in the north-eastern Gulf of Mexico provides a unique opportunity to test these systems down to scales previously unobtainable. In this study, background error covariance assumptions in the 3DVar assimilation process are perturbed to understand the effect on the solution relative to the withheld dense drifter data. Results show that the amplitude of the background error covariance is an important factor as expected, and a proposed new formulation provides added skill. In addition, the background error covariance time correlation is important to allow satellite observations to affect the results over a period longer than one daily assimilation cycle. The results show the new background error covariance formulations provide more accurate placement of frontal positions, directions of currents and velocity magnitudes. These conclusions have implications for the implementation of 3DVar systems as well as the analysis interval of 4DVar systems. Published by Elsevier Ltd.

Ocean model fields are being routinely used for forecasting the spreading of pollutants, oil spills, and for biogeochemical transport. Recent observations and advances in our understanding of ocean processes indicate there is an explosion of flow instabilities in the submesoscale range. While submesoscale flows have a significant impact on transport at their own scales, they require much more extensive data sets and numerical computations. Therefore, transport carried out by submesoscale flows is quite challenging to approach deterministically. In this study, we put forward a hybrid approach by combining deterministic Lagrangian coherent structures (LCS) to compute transport over the mesoscale range with statistical Lagrangian subgridscale (LSGS) models for the underresolved submesoscale motions. We apply this approach to particle transport in the Gulf Stream region, which exhibits indications of submesoscale activity from both models and observations. We consider HYCOM solutions at two resolutions. In the 1/12° computation, mesoscale features are well resolved but submesoscales are not resolved, while the 1/48° computation captures some of the submesoscale flow instabilities as well. By using metrics of relative dispersion, we investigate three LSGS models and demonstrate that they can be useful in correcting the underestimation of submesoscale dispersion in the 1/12° solution, with respect to relative dispersion obtained from the 1/48° solution and an observational result.

Relative dispersion in the Liguro-Provencal basin (a subregion of the Mediterranean Sea) is investigated using clusters of surface drifters deployed during two Marine Rapid Environment Assessment (MREA) experiments covering different months in 2007 and 2008, respectively. The clusters have initial radii of less than 1 km, or an order of magnitude below a typical deformation radius (approximately 10–20 km). The data set consists of 45 original pairs and more than 50 total pairs (including chance ones) in the spatial range between 1 and 200 km. Relative dispersion is estimated using the mean square separation of particle pairs and the Finite Scale Lyapunov Exponents (FSLEs). The two metrics show broadly consistent results, indicating in particular a clear exponential behaviour with an e-folding time scale between 0.5 and 1 days, or Lyapunov exponent λ in the range of 0.7–1 days−1. The exponential phase extends for 4–7 days in time and between 1 and 10–20 km in separation space. To our knowledge, this is only the third time that an exponential regime is observed in the world ocean from drifter data. This result suggests that relative dispersion in the Liguro-Provencal basin is nonlocal, namely controlled mainly by mesoscale dynamics, and that the effects of the sub-mesoscale motions are negligible in comparison. NCOM model results are used to complement the data and to quantify errors arising from the sparse sampling in the observations.

A method for the reconstruction of the mesoscale Eulerian velocity field based on Lagrangian data at given sampling period is presented. A variational approach is used, where information on the float positions are combined with a simple model constraint describing the motion of particles advected in a velocity field. The velocity field of a priori specified space scale is estimated minimizing a cost function which measures the distance between the observed float positions and the model predicted ones for each sampling period. The method is first implemented considering the time-independent approximation of the velocity correction during a time interval shorter than the typical time scale of the mesoscale field. In a second step, this approximation is relaxed to consider inertial oscillations superimposed on the mesoscale field. The method is tested on a numerical regional circulation on the Northwestern Mediterranean Sea, characterized by an Eulerian time scale significantly longer than the inertial period. The observational coverage of the region varies between 80% and 16% and the corresponding error in the velocity reconstruction varies from less than 10% to approximately 50%, with a sampling period of the day. For smaller sampling intervals, the reconstruction is further improved in the time-dependent approximation. Finally, the method is used to combine information from Lagrangian data and General Circulation Models. The results suggest the useful application of the method for assimilation studies.

Results from visual sightings of large floating debris are presented, taken in the Ligurian Sea, a sub-basin of the north-western Mediterranean Sea which belongs to the recently stated ‘‘Cetacean Sanctuary’’. Data have been collected during three oceanographic cruises, during the summer of 1997 and 2000. Results for the 1997 data suggest a debris density of the order of 15–25 objects km � 2 , while for the 2000 data, a lower density of the order of 3–1.5 objects km � 2 is found. The difference between the two results appears statistically significant using simple tests. Possible reasons for the observed variability are discussed, including meteorological forcing, marine currents and debris input variability. 2003 Elsevier Ltd. All rights reserved.

The predictability of Lagrangian particle trajectories in the Adriatic Sea (a semi-enclosed sub-basin of the Mediterranean Sea) over a period of 1-2 weeks is investigated using three clusters consisting of 5-7 drifters. The analysis is conducted using a Gauss-Markov Lagrangian particle model, which relies on the estimate of climatological mean flow field, persistence of turbulence, and assimilation of velocity data from the surrounding drifters through a Kalman filtering technique. The results are described using the data density N R defined as the number of drifters within a distance on the order of the Rossby radius of deformation from the particle to be predicted. The clusters are inherently different with respect to this characteristic property with values ranging from N R

A model/data comparison was performed between simulated drifters from a high-resolution numerical simulation of the North Atlantic and a data set from in situ surface drifters. The comparison makes use of pseudo-Eulerian statistics such as mean velocity and eddy kinetic energy, and Lagrangian statistics such as integral time scales. The space and time distribution of the two data sets differ in the sense that the in situ drifters were released inhomogeneously in space and time while the simulated drifters were homogeneously seeded at the same time over a regular 1° grid. Despite this difference, the total data distributions computed over the complete data sets show some similarities that are mostly related to the large-scale pattern of Ekman divergence/convergence. Comparisons of eddy kinetic energy and root mean square velocity indicate that the numerical model underestimates the eddy kinetic energy in the Gulf Stream extension and in the ocean interior. In addition, the model Lagrangian time scales are longer in the interior than the in situ time scales by approximately a factor of 2. It is suggested that this is primarily due to the lack of high-frequency winds in the model forcing, which causes an underestimation of the directly forced eddy variability. Regarding the mean flow, the comparison has been performed both qualitatively and quantitatively using James' statistical test. The results indicate that over most of the domain, the differences between model and in situ estimates are not significant. However, some areas of significant differences exist, close to high-energy regions, notably around the Gulf Stream path, which in the model lies slightly north of the observed path, although its strength and structure are well represented overall. Mean currents close to the buffer zones, primarily the Azores Current, also exhibit significant differences between model results and in situ estimates. Possibilities for model improvement are discussed in terms of forcings, buffer zone implementations, turbulence and mixed layer parameterizations, in light of our model/data comparison.

In this paper, the statistical properties of the mean flow reconstruction using Lagrangian data are studied, considering the classical “binning” approach based on space-time averaging of finite difference velocity estimates. The work is performed numerically, using as the test flow a solution from a high resolution MICOM simulation of the North Atlantic. A set of trajectories are computed, simulating the motion of surface drifters initially launched on a regular 1°×1° array, transmitting positions every Δt=12 h, and analyzed over approximately 2 years of the simulation. The drifter distribution in time is influenced by the Ekman flow, resulting in maximum data concentration in the subtropical convergence regions and minimum concentration in the upwelling regions. Pseudo-Eulerian averages UpE, computed from Langrangian data, are compared to “true” Eulerian averages UE, computed from grid point velocities inside 1°×1° bins for approximately 2 years. For the full Lagrangian data set (which is substantially larger than the WOCE requirement), UpE−UE is on the order of 10–20 cm/s in regions of major ocean currents. These differences are usually not significant with respect to the sampling error, due to subgrid-scale variability and finite sampling, except in a few regions. Patterns of the magnitude of the differences between UpE and UE in these regions show that UpE tends to underestimate (overestimate) the velocity in the eastern equatorial upwelling regime/South Equatorial current (western boundary currents). This study suggests that these under/overestimates by pseudo-Eulerian averaging of Lagrangian data are related to a bias due to mesoscale divergences, and result in nonzero correlations between instantaneous drifter concentration and velocity, UB=〈u′c′〉/C Davis, 1998 , Gent and McWilliams, 1990 . In this framework, the overestimates (underestimates) are interpreted as due to preferential (reduced) sampling of high velocity regions by Lagrangian particles, due to convergent (divergent) phenomena. A similar phenomenon has been observed for real drifters and biological organisms. The overestimates are found to increase with sub-sampling in space and decrease with sub-sampling in time. For Δt=3 days, we actually find underestimates, probably because instantaneous high velocities are smoothed and energetic drifters are not appropriately accounted for in the bins. Direct implications of the results for the analysis of real data, and directions for future work (in particular investigation of the bias) are discussed.

The dispersion processes in semi-enclosed basins with stationary flows have been studied by means of an Eulerian advection–diffusion model. Special interest has been given to quantities which characterize the macroscopic state of the system: the normalized total quantity C(t) of a tracer inside the basin, and its residence time T. The sensitivity of the results to a number of parameters, including flow structure, Peclet number, initial and boundary conditions, have been studied. In particular, the difference between unidirectional flows and flows with recirculations has been investigated. A priori estimates of the value of T are provided, valid for a wide range of parameters. For unidirectional flows, T≃Tout+τ0, where Tout is the time a substantial amount of tracer takes to reach the outflow and τ0 is the inverse of the principal eigenvalue of the advection–diffusion operator. Both parameters Tout and τ0 can be computed knowing the flow structure and the Peclet number, and they are representative, respectively, of the initial and asymptotic phase of dispersion. For flows with recirculations it is found that T≃τ0 over the whole parameter space considered. Conceptually, the difference is due to the tendency of the recirculation to trap the tracer, which enhances the importance of the asymptotic phase. For these flows, the values of T do not change significantly in the range of realistic Peclet numbers considered, suggesting that the macroscopic dispersion properties depend only weakly on the eddy diffusion coefficient. This can have obvious consequences in terms of the possibility of a priori estimates of the residence time, even without having available accurate measurements of the diffusivity.

We study a discrete form of the barotropic vorticity equation used in the numerical studies of the ocean general circulation. Our goal is to contribute to the understanding of the numerical solutions, introducing some new tools of investigation. We begin by examining the basic properties of the discrete solutions for time-independent forcing. All the solutions are shown to be ultimately bounded in a finite region in the phase space and an estimate of the energy and enstrophy bounds is made. Steady-state solutions are shown always to exist for abitrary forcing and non-zero dissipation. A sufficient condition of asymptotic stability and uniqueness for the steady-state solutions is derived. The limit solutions for zero dissipation and infinite forcing are considered. For the infinite forcing case, a class of asymptotic steady-state solutions φk is recovered, proportional to the eigenvectors of the Laplacian operator. In this limit, the discrete problem is shown to be different from the continuous problem that allows for a wider class of asymptotic solutions. The analytical results have been used to study a number of numerical solutions computed for the single-gyre and double-gyre wind stresses and for various values of the forcing and dissipation parameters. We focus on solutions at high non-linearity, obtained either for decreasing dissipation or for increasing forcing. When the dissipation is decreased, the energy of the solutions increases at a slower rate than the upper bound. For sufficiently small values of the dissipation parameters, the steady-state solutions become unstable and the computed solutions are time dependent. The transition to instability is characterized by the rapid increase of the non-linear contribution to the Jacobian matrix. When the forcing is increased, the energy of the solutions increases and tends to the theoretical bound. These solutions are always stationary and asymptotically stable. The non-linear contribution to the Jacobian matrix increases linearly with the forcing. The steady-state solutions in this case converge to the asymptotic solutions φ1andφ2 for the single-gyre and the double-gyre respectively.

A complete and systematic derivation of Eulerian variational principles in fluid dynamics is given; the original Lagrangian, describing the fluid as a continuous system of particles, is systematically transformed by canonical transformations. Some of the variational principles known in the literature are recovered, and a new formulation is found. The formulations are discussed with respect to their applications in fluid dynamics.