Zulema D. Garraffo

Evaluating the Decomposition of Tropical Atlantic Drifter Observations

Abstract Because the tropical Atlantic is characterized by regions of strong seasonal variability that have been sampled inhomogeneously by surface drifters, Eulerian averages of these Lagrangian observations in spatially fixed bins may be aliased. In the Pacific, this problem has been circumvented by first calculating seasonal or monthly means. In the Atlantic, such an approach is of limited value because of the relatively sparse database of drifter observations. As an alternative, a methodology is developed in which drifter-observed currents and sea surface temperatures are grouped into bins and, within each bin, simultaneously decomposed into a time-mean, annual and semiannual harmonics, and an eddy residual with a nonzero integral time scale. The methodology is evaluated using a temporally homogeneous SST product and in situ SST observations, and also using simulated drifter observations in an eddy-resolving model of the Atlantic Ocean. These analyses show that, compared to simple bin averaging, the d...

Seasonality of the submesoscale dynamics in the Gulf Stream region

Frontogenesis and frontal instabilities in the mixed layer are known to be important processes in the formation of submesoscale features. We study the seasonality of such processes in the Gulf Stream (GS) region. To approach this problem, a realistic simulation with the Hybrid Coordinate Ocean Model is integrated for 18 months at two horizontal resolutions: a high-resolution (1/48°) simulation able to resolve part of the submesoscale regime and the full range of mesoscale dynamics, and a coarser resolution (1/12°) case, in which submesoscales are not resolved. Results provide an insight into submesoscale dynamics in the complex GS region. A clear seasonal cycle is observed, with submesoscale features mostly present during winter. The submesoscale field is quantitatively characterized in terms of deviation from geostrophy and 2D dynamics. The limiting and controlling factor in the occurrence of submesoscales appears to be the depth of the mixed layer, which controls the reservoir of available potential energy available at the mesoscale fronts that are present most of the year. Atmospheric forcings are the main energy source behind submesoscale formation, but mostly indirectly through mixed layer deepening. The mixed layer instability scaling suggested in the (Fox-Kemper et al., J Phys Oceanogr 38:1145–1165, 2008) parametrization appears to hold, indicating that the parametrization is appropriate even in this complex and mesoscale dominated area.

Lagrangian spin parameter and coherent structures from trajectories released in a high-resolution ocean model

A study of the mesoscale eddy field in the presence of coherent vortices, by means of Lagrangian trajectories released in a high-resolution ocean model, is presented in this paper. The investigation confirms previous results drawn from real float data statistics (Veneziani et al., 2004) that the eddy field characteristics are due to the superposition of two distinct regimes associated with strong coherent vortices and with a typically more quiescent background eddy flow. The former gives rise to looping trajectories characterized by subdiffusivity properties due to the trapping effect of the vortices, while the latter produces nonlooping floats characterized by simple diffusivity features. Moreover, the present work completes the study by Veneziani et al. (2004) in regard to the nature of the spin parameter , which was used in the Lagrangian stochastic model that best described the observed eddy statistics. The main result is that the spin obtained from the looping trajectories not only represents a good estimate of the relative vorticity of the vortex core in which the loopers are embedded, but it is also able to follow the vortex temporal evolution. The Lagrangian parameter is then directly connected to the underlying Eulerian structure and could be used as a proxy for the relative vorticity field of coherent vortices.

Lagrangian data in a high-resolution numerical simulation of the North Atlantic: II. On the pseudo-Eulerian averaging of Lagrangian data

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° X 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 U pE , computed from Langrangian data, are compared to true Eulerian averages U E , computed from grid point velocities inside 1° X 1° bins for approximately 2 years. For the full Lagrangian data set (which is substantially larger than the WOCE requirement), U pE - U E 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 U pE and U E in these regions show that U pE 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, U B = /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 Δ t = 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.

Parameterizations of Lagrangian spin statistics and particle dispersion in the presence of coherent vortices

Coherent vortices are known to play an important role in transport processes of ideal flows such as two-dimensional and quasi-geostrophic turbulent flows. In this paper, their effect on eddy dispersion and diffusivity is studied in a realistic oceanic flow, using synthetic Lagrangian data simulated within a high-resolution ocean general circulation model in the Gulf Stream recirculation region. The possibility of using a Lagrangian Stochastic Model (LSM) with nonzero mean spin statistic, ,t o parameterize the observed characteristics is considered. The probability distribution of the parameter (which is representative of the looping behavior of trajectories embedded inside the coherent vortices) is also studied. The main result is that the LSM with a tri-modal -distribution is able to reproduce the eddy diffusivity and dispersion characteristics, especially at short and intermediate time scales. Particularly well predicted are the effects of the coherent vortices, which are the enhanced spreading of particles at short times due to the high eddy energy of the vortices and the inhibited diffusion at intermediate times due to the vortex trapping mechanisms. The LSM parameterization of eddy dispersion is shown to be more appropriate than the commonly used asymptotic eddy-diffusivity approximation. More complex than tri-modal distributions of mean spin are also considered in the LSM, and they are found to produce only slight changes in the predictions of eddy statistics and dispersion.

Lagrangian Velocity Distributions in a High-Resolution Numerical Simulation of the North Atlantic

The statistical properties of Lagrangian velocities in a high-resolution numerical simulation of the North Atlantic Ocean are analyzed and discussed in the framework of particle dispersion parameterizations. Consistent with previous analyses of float trajectories, the modeled velocity distribution is shown to be non-Gaussian, both at the surface and at 1500 m. These results can have significant implications on oceanographic research, as they suggest that current parameterizations of particle dispersion by linear stochastic processes or eddy-diffusivity approaches may be incorrect, since they assume Gaussian velocity distributions. The results also indicate the need for empirical parameterizations of particle dispersion based on nonlinear stochastic processes. It is shown that, even for a truly non-Gaussian dataset, a Gaussian probability distribution function can be spuriously recovered when the sampling density is too low. The best compromise between data sampling and space averaging when a limited amount of data are available, as is the case in most field observations, is then identified.

Pathways and variability at intermediate depths in the tropical Atlantic

Abstract Oceanographie and meteorological data as well as model results are analyzed to study the pathways and the temporal variability of the intermediate depth (800-1100 m) flow in the tropical Atlantic (9°S to 7°N). The mean flow is dominated by zonal currents which interact with the western boundary current. These currents frequently experience reversals of the zonal and meridional flow. The primary focus in the analysis of the variability is on the region around 6°S. The observations reveal temporal variability on mesoscale, annual and interannual time scales. Several westward propagating signals can be identified, with propagation velocities between 5 and 7 cm s 1 . Two zonal length scales (500-700 km and more than 2000 km) are observed. It is hypothesized that these are due to planetary waves. A comparative analysis of observations and model velocities reveals striking similarities in their time and length scales. Sample spectra of the model velocities show a dominant peak of the spectral energy density at a wave length between 500 km and 1100 km. Additionally, a longer wave with a zonal wave length of about 5000 km is present, which can not be resolved by the spectral analysis. In the time space the spectral analysis for the zonal and meridional velocity reveals coinciding peaks at periods of 45 days, 66 days and one year. For the latter two periods the energy for the two velocity components are quite similar. An analytical planetary wave solution shows that a superposition of a mesoscale and an annual planetary wave is sufficient to reproduce a large part of the variability found in the observations and the model. The wave with an annual period is most likely due to the annual cycle of the wind field.

Analysis of a general circulation model product: 1. Frontal systems in the Brazil/Malvinas and Kuroshio/Oyashio regions

In the present paper (part 1 of 2), the product of the Semtner and Chervin general circulation model (GCM) is compared with available observations in the frontal areas of the Brazil/Malvinas and the Kuroshio/Oyashio confluences. The dimensionality of the systems studied is reduced by using the empirical orthogonal functions (EOF) and frontal density methods. The two sets of data utilized to validate the model are the sea surface temperature (SST) from the satellite observations and temperature fields product from the GCM at levels 1 (12.5 m), 2 (37.5 m) and 6 (160 m). Comparisons are made between the dominant empirical modes and the locus of maximum probability for observations and model product. The model reproduces intense thermal fronts at the surface and in the upper layers. In the upper layer (level 1) they are induced by the internal dynamics of the model and not by the restoring of the model to climatology alone. The variability of these fronts is less pronounced in the model than in the observations. The dominant period in the observations is annual with contributions of semiannual and high frequency oscillations. In the model, the dominant variability is also annual at all analyzed levels. A semiannual oscillation contributed to a lower degree and is related to eddies that, in the model, have an annual and semiannual periodicity. For the regions examined, the location of the fronts are reproduced in the model within differences of 4° to 5° with observations. In the Brazil/Malvinas region, the Confluence front is reproduced approximately 4° towards the west of the observed front. This appears to be due to the resolution of the model that, in a 0.5° × 0.5° grid, does not resolve the sharp slope at the edge of the Argentine continental shelf. The maximum southward penetration of the warm tongue of Brazil waters occurs in the model approximately 4° towards the north. This is related to the fact that, in the model, the Malvinas transport doubles the one derived from the observations. This might be due to the effect of a large modeled transport for the Circumpolar Current or, again, to a poor resolution of the topography. In the Kuroshio area, the Oyashio front, which in observations is more pronounced at the surface than in the lower layers, is well reproduced in the surface temperature field. On the contrary, the Kuroshio front, more intense in the lower layers but still marked in the satellite observations, is visible in the model only below 160 m. The front is not present in the surface temperature field but, as a consequence of the thermal wind balance, an intense eastward flow at the location of the Kuroshio Extension is observed in the model velocity field. When compared with the observations, the location of the Extension is shifted approximately 5° towards the south. This indicates a shift in latitude between the modeled and observed latitude of separation. The resolution of the model is marginal to reproduce the process of eddy formation, but large scale eddies are observed in the model in both analyzed areas. They are generated as a pinch of the main flow with an annual and semiannual periodicity. We conclude that some of the differences between model and observations, like the differences found in the locations of the fronts, and the diminished variability, will decrease with a higher resolution.

Global Ocean Prediction Using HYCOM

One important aspect of ocean model design is the choice of the vertical coordinate system. Traditional ocean models use a single coordinate type to represent the vertical, but model comparison exercises performed in Europe (DYnamics of North Atlantic MOdels - DYNAMO) (Willebrand et al., 2001) and in the United States (Data Assimilation and Model Evaluation Experiment - DAMEE) (Chassignet et al., 2000) have shown that none of the three main vertical coordinates presently in use (depth [z-levels], density [isopycnal layers], or terrain-following [sigma-levels]) can by itself, be optimal everywhere in the ocean. The HYbrid Coordinate Ocean Model (HYCOM) (Bleck, 2002) is configured to combine all three of these vertical coordinate types. It is isopycnal in the open, stratified ocean, but uses the layered continuity equation to make a dynamically smooth transition to a terrain-following coordinate in shallow coastal regions, and to z-level coordinates in the mixed layer and/or unstratified seas. The hybrid coordinate extends the geographic range of applicability of traditional isopycnic coordinate circulation models toward shallow coastal seas and unstratified parts of the world ocean. It maintains the significant advantages of an isopycnal model in stratified regions while allowing more vertical resolution near the surface and in shallow coastal areas, hence providing a better representation of the upper ocean physics

Analysis of a general circulation model: 2. Distribution of kinetic energy in the South Atlantic and Kuroshio/Oyashio systems

The energy of the model transient eddies at 37.5 m is compared with Geosat altimeter observations, for the South Atlantic Ocean and for the Kuroshio system. The model shows areas of transient motions overlapping the ones obtained from Geosat altimeter data. For the South Atlantic Ocean, the modeled eddy kinetic energy is smaller than the one observed with Geosat, by a factor of 3 for area average on the whole South Atlantic region, and by a factor of 4 for its western boundary. On the Agulhas system, transient eddy activity develops in the region where the Agulhas current retroflects. In the western South Atlantic, the modeled eddy activity is concentrated on the Confluence front; observed variability along a more extended region following the topography is not resolved in the model. For the Kuroshio region, the energy level of the modeled transient motions is comparable with Geosat observations, but the model eddy activity is more concentrated in the Kuroshio Current and not in the Kuroshio extension. The observations show the opposite. For the South Atlantic Ocean, a comparison is also done between model eddy kinetic energy (defined as including standing and transient eddy contributions) with values obtained from surface drifters. The analysis shows differences in the western boundary, and good agreement across the South Atlantic Ocean between 35°S and 45°S. In this formulation, the model mean energy level is smaller than the observed with drifters from the First GARP Global Experiment; differences might be due to an overestimation in the values obtained with the drifters.