• Aref, H., 1984: Stirring by chaotic advection. J. Fluid Mech., 143 , 121.

  • Artegiani, A., , Bregant D. , , Paschini E. , , Pinardi N. , , Raicich F. , , and Russo A. , 1997: The Adriatic Sea general circulation. Part I: Air–sea interactions and water mass structure. J. Phys. Oceanogr., 27 , 14921514.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. E., 1985: Drifter observation of coastal currents during CODE. The method and descriptive view. J. Geophys. Res., 90 , 47414755.

  • Doglioli, A. M., , Griffa A. , , and Magaldi M. G. , 2004: Numerical study of a coastal current on a steep slope in presence of a cape: The case of the Promontorio di Portofino. J. Geophys. Res., 109 .C12033, doi:10.1029/2004JC002422.

    • Search Google Scholar
    • Export Citation
  • Dunlap, E., , Tang C. C. L. , , and Wang C. K. , 2004: Comparison of model surface currents and drifter data from the Grand Banks. Canadian Hydrography and Ocean Science Tech. Rep. 236, Department of Fisheries and Oceans Canada, Dartmouth, NS, Canada, 34 pp.

  • Dutkiewicz, S., , Griffa A. , , and Olson D. B. , 1993: Particle diffusion in a meandering jet. J. Geophys. Res., 98 , 1648716500.

  • Falco, P., , Griffa A. , , Poulain P-M. , , and Zambianchi E. , 2000: Transport properties in the Adriatic Sea as deduced from drifter data. J. Phys. Oceanogr., 30 , 20552071.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Griffa, A., , Piterbarg L. I. , , and Ozgokmen T. , 2004: Predictability of Lagrangian particle trajectories: Effects of smoothing of the underlying Eulerian flow. J. Mar. Res., 62 , 135.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hopkins, T. S., , Artegiani A. , , Kinder C. , , and Pariante R. , 1999: A discussion of the northern Adriatic circulation and flushing as determined from the ELNA hydrography. The Adriatic Sea, Ecosystem Rep. EUR 18834, Vol. 32, European Commission, 85–106.

  • Kuznetsov, L., , Toner M. , , Kirwan A. D. Jr., , Jones C. K. R. T. , , Kantha L. H. , , and Choi J. , 2002: The Loop Current and adjacent rings delineated by Lagrangian analysis of the near-surface flow. J. Mar. Res., 60 , 405429.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paris, C. B., , Cowen R. K. , , Claro R. , , and Lindeman K. C. , 2005: Larval transport pathways from Cuban snapper (Lutjanidae) spawning aggregations based on biophysical modeling. Mar. Ecol. Prog. Ser., 296 , 93106.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poje, A. C., , and Haller G. , 1999: Geometry of cross-stream mixing in a double-gyre ocean model. J. Phys. Oceanogr., 29 , 16491665.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., 1999: Drifter observations of surface circulation in the Adriatic Sea between December 1994 and March 1996. J. Mar. Syst., 20 , 231253.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., 2001: Adriatic Sea surface circulation as derived from drifter data between 1990 and 1999. J. Mar. Syst., 29 , 332.

  • Poulain, P-M., , Barbanti R. , , Cecco R. , , Fayos C. , , Mauri E. , , Ursella L. , , and Zanasca P. , cited. 2004a: Mediterranean surface drifter database: 2 June 1986 to 11 November 1999. OGS Tech. Rep. 78/2004/OGA/31. [Available online at http://poseidon.ogs.trieste.it/drifter/database_med.].

  • Poulain, P-M., , Mauri E. , , and Ursella L. , 2004b: Unusual upwelling event and current reversal off the Italian Adriatic coast in summer 2003. Geophys. Res. Lett., 31 .L05303, doi:10.1029/2003GL019121.

    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., , Barbanti R. , , Ursella L. , , and Cecco R. , cited. 2006: Adriatic surface drifter database: 21 September 2002 to 29 February 2004. OGS Tech. Rep. [Available online at http://poseidon.ogs.trieste.it/doga/sire/dolcevita/database_ dolcevita/.].

  • Samelson, R. M., 1996: Chaotic transport by mesoscale motions. Stochastic Modeling in Physical Oceanography, R. J. Aller, P. Miller, and B. Rozovskii, Eds., Birkhauser, 423–438.

    • Search Google Scholar
    • Export Citation
  • Spaulding, M. L., , Isaji T. , , Allen A. , , Hall P. , , and Howelett E. , 2005: Application of a random flight model in search and rescue model systems. Third Lagrangian Analysis and Prediction of Coastal and Ocean Dynamics (LAPCOD) Meeting, Lerici, Italy, ONR and Cosponsors. [Available online at http://www.rsmas.miami.edu/LAPCOD/2005-Lerici/.].

  • Thompson, K. R., , Sheng J. , , Smith P. C. , , and Cong I. , 2003: Prediction of surface currents and drifter trajectories on the inner Scotian Shelf. J. Geophys. Res., 108 .3287, doi:10.1029/2001JC001119.

    • Search Google Scholar
    • Export Citation
  • Thomson, D. J., 1987: Criteria for the selection of stochastic models of particle trajectories in turbulent flows. J. Fluid Mech., 210 , 113153.

    • Search Google Scholar
    • Export Citation
  • Ursella, L., , Poulain P-M. , , and Signell R. , 2006: Surface drifter derived circulation in the northern and middle Adriatic Sea: Response to wind regime and season. J. Geophys. Res., 111 .C03S04, doi:10.1029/2005JC003177.

    • Search Google Scholar
    • Export Citation
  • Wiggins, S., 1992: Chaotic Transport in Dynamical Systems. Springer, 301 pp.

  • Zore-Armanda, M., 1956: On gradient currents in the Adriatic Sea. Acta Adriat., 8 , 6. 138.

  • Zore-Armanda, M., 1969: Water exchange between the Adriatic and the Eastern Mediterranean. Deep-Sea Res., 16 , 171178.

  • View in gallery

    Description of geographic locations in the Adriatic Sea. The topography of the area is also shown (with 50-m contour levels).

  • View in gallery

    (top) Mean flow obtained by averaging the historical drifter velocities over 0.1° square bins. Gray vectors are for bins with more than 5 but less than 10 independent measurements, n* (where n* is computed considering a Lagrangian decorrelation time scale TL = 2 days); black vectors are for bins with more than 10 independent measurements. The standard error ellipses are computed with respect to the major and minor axis of variability. Also shown are the 100- and 160-m isobaths. (bottom) EKE field obtained by averaging the Lagrangian eddy velocities over 0.1° square bins. Only results for bins with n* ≥ 5 are shown. The thick black lines identify the DART area of interest, while the thin lines identify the following 5 subregions: POLYN (north polygon), POLYS (south polygon), GARGN (north Gargano Cape), GARGS (south Gargano Cape), and RECIRC (Gargano recirculation).

  • View in gallery

    Forward and backward trajectories of the drifters from/to (top) POLYN, (middle) GARGN, and (bottom) RECIRC; the corresponding subregions are colored in gray for clarity. Both tforw and tback are equal to 10 days.

  • View in gallery

    (top) Initial positions of the POLYN and POLYS drifters. Filled (open) circles identify the BCint (BCext) drifters. (bottom) Histograms of the arrival time taken by the ngarg drifters (see Table 2) to reach the Gargano Cape from POLYN and POLYS.

  • View in gallery

    Maps of drifter concentration computed over a 0.2° × 0.2° spatial grid at various fixed times tforw for the BCint drifters from the POLYN region. The initial number of these drifters is 58. See text for percentage definitions.

  • View in gallery

    As in Fig. 5, but relative to the BCext drifters from POLYN (their initial total number is 33).

  • View in gallery

    Initial positions of the BCint and BCext drifters in (top) GARGN, and initial positions of the total drifters in (middle) GARGS and (bottom) RECIRC.

  • View in gallery

    As in Fig. 5, but relative to the sum of the BCint and BCext drifters from GARGN (their initial total number is 46).

  • View in gallery

    As in Fig. 5, but relative to the total drifters from GARGS (their initial number is 104).

  • View in gallery

    As in Fig. 5, but relative to the total drifters from RECIRC (their initial number is 23).

  • View in gallery

    Mean flow obtained by averaging the drifters velocities over 0.1° square bins for the four seasonal periods. Gray vectors are for bins with more than 3 but less than 10 independent measurements; black vectors are for bins with more than 10 independent measurements.

  • View in gallery

    (left) Forward and (right) backward trajectories relative to the POLYN subregion, during the four different seasons. The time intervals tforw and tback are equal to 10 days.

  • View in gallery

    As in Fig. 5, but relative to the total drifters from POLYN (Bcint + BCext), during two extended seasonal periods: the (left) fall–winter and (right) spring–summer periods.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 21 21 3
PDF Downloads 3 3 0

Historical Drifter Data and Statistical Prediction of Particle Motion: A Case Study in the Central Adriatic Sea

View More View Less
  • 1 RSMAS/MPO, University of Miami, Miami, Florida
  • | 2 RSMAS/MPO, University of Miami, Miami, Florida, and ISMAR/CNR, La Spezia, Italy
  • | 3 OGS, Trieste, Italy
© Get Permissions
Full access

Abstract

In this paper, a method to analyze historical surface drifter data is presented that is aimed at investigating particle evolution as a function of initial conditions. Maps of drifter concentration at different times are built and interpreted as maps of the probability of finding a particle at a given time in the neighborhood of a given point in the domain. A case study is considered in a coastal area of the middle Adriatic Sea (a subbasin of the Mediterranean Sea) around the Gargano Cape, which is the focus of a newly planned experiment, the Dynamics of the Adriatic in Real Time (DART). A specific application is considered that seeks to improve the DART Lagrangian sampling planning.

The results indicate that the analysis of historical drifters can provide very valuable information on statistical particle prediction to be used in experiment design. In the DART region, particle dynamics appear mostly controlled by the upstream properties of the boundary current as well as by the presence of a stagnation point located offshore of the tip of Gargano and separating two cross-basin recirculations. A significant seasonal dependence is observed, with drifters being more likely to leave the boundary current in winter and fall, when the current is wider and more connected to the cross-basin recirculations. Future developments are discussed, including joint analyses with numerical model results.

* Current affiliation: Ocean Sciences Department, University of California, Santa Cruz, Santa Cruz, California

Corresponding author address: Dr. Milena Veneziani, Ocean Sciences Department, University of California, Santa Cruz, Santa Cruz, CA 95064. Email: milenav@pmc.ucsc.edu

Abstract

In this paper, a method to analyze historical surface drifter data is presented that is aimed at investigating particle evolution as a function of initial conditions. Maps of drifter concentration at different times are built and interpreted as maps of the probability of finding a particle at a given time in the neighborhood of a given point in the domain. A case study is considered in a coastal area of the middle Adriatic Sea (a subbasin of the Mediterranean Sea) around the Gargano Cape, which is the focus of a newly planned experiment, the Dynamics of the Adriatic in Real Time (DART). A specific application is considered that seeks to improve the DART Lagrangian sampling planning.

The results indicate that the analysis of historical drifters can provide very valuable information on statistical particle prediction to be used in experiment design. In the DART region, particle dynamics appear mostly controlled by the upstream properties of the boundary current as well as by the presence of a stagnation point located offshore of the tip of Gargano and separating two cross-basin recirculations. A significant seasonal dependence is observed, with drifters being more likely to leave the boundary current in winter and fall, when the current is wider and more connected to the cross-basin recirculations. Future developments are discussed, including joint analyses with numerical model results.

* Current affiliation: Ocean Sciences Department, University of California, Santa Cruz, Santa Cruz, California

Corresponding author address: Dr. Milena Veneziani, Ocean Sciences Department, University of California, Santa Cruz, Santa Cruz, CA 95064. Email: milenav@pmc.ucsc.edu

1. Introduction

Predicting particle motion in the ocean is of great importance for a number of practical applications. These applications include prediction of pollutant spreading, oil spill containment, and search-and-rescue activities, as well as assessment of biological quantities such as larvae spreading, or warfare exercises like mine tracking.

Particle prediction is a very challenging problem because of the well-known conceptual reason that trajectories are extremely sensitive to the details of the advective Eulerian velocity field and are highly chaotic (e.g., Aref 1984; Samelson 1996). It is therefore not surprising that, despite the great advances in ocean circulation models in the last two decades, quantitatively predicting the motion of single particles remains a highly complex and demanding problem (e.g., Dunlap et al. 2004; Thompson et al. 2003). Even small errors in the estimation of ocean currents can drastically change particle trajectories (Griffa et al. 2004). Furthermore, sources of errors in modeling ocean currents are ubiquitous due to uncertainties in the wind forcing, bathymetry, coastline, and internal stratification, and the model resolution and subgrid-scale parameterizations. Similar difficulties are found when particle prediction is performed on the basis of extensive velocity datasets, such as current maps constructed with high-frequency coastal radars (J. O’Donnell et al. 2006, personal communication; Spaulding et al. 2005). Uncertainties in the observed velocity and the presence of unresolved scales of motion can lead to significant errors in particle prediction. While it can be expected that future improvements in both models and measurements and in their combined use through data assimilation will lead to significant enhancement of particle prediction, the present state of the art is still problematic.

Given the difficulties in deterministically predicting the motion of single particles, in practical applications a statistical point of view is often adopted. Instead of trying to reproduce the exact motion of single trajectories, information is sought on the probability of finding a particle in a specific region and at a given time. In the framework of numerical modeling, this problem has been treated using two main approaches. The “stochastic” approach (e.g., Thomson 1987; Dutkiewicz et al. 1993; Paris et al. 2005) consists of perturbing the numerical trajectories predicted by a circulation model using a random noise, which represents the unresolved motion and is parameterized through a Lagrangian stochastic model (LSM). An ensemble of particle trajectories is generated for each initial condition, and the ensemble evolution in space and time provides information on the evolution of the probability density function (PDF). The “dynamical system” approach (e.g., Wiggins 1992; Poje and Haller 1999; Kuznetsov et al. 2002), on the other hand, focuses on the general Lagrangian structure of the flow generated by a circulation model, identifying its hyperbolic points and manifolds, which tend to act as transport barriers. Ensembles of initial conditions (ICs) can be identified and are characterized by similar Lagrangian characteristics or transport fates. Both the stochastic and the dynamical system approaches are promising, but they still present a number of open questions. In the former, questions remain on the identification of the correct LSM and of its parameter values in order to effectively parameterize the unresolved motion. In the latter, development is still needed in the case of realistic flows with high transients, where significant geometrical properties might be difficult to identify and characterize.

In this paper, we propose an approach to the problem of statistical particle prediction based on the analysis of in situ historical drifter data. The methodology can be considered complementary to the mentioned modeling approaches, and is applied to a specific case study in the central Adriatic Sea, a subbasin of the Mediterranean Sea, where a significant historical drifter dataset has been collected over the years (e.g., Poulain 2001). We focus on the region of a new experiment that is presently being planned, the Dynamics of the Adriatic in Real Time (DART; see information online at http://oceans.deas.harvard.edu/haley/DART05). The DART region is in the coastal area of the middle Adriatic, close to the Gargano Cape. It is an area with strong topographic control and significant mesoscale activity related to instabilities of the boundary current. The DART experiment has an important Lagrangian component and therefore provides an additional and more specific motivation to the present study. We expect to use our results as guidance for the planning of the experimental Lagrangian sampling strategy.

The goal of the paper is to describe a methodology and its application to a coastal region. We aim at addressing the following question: given a particle launched at point x0, what is the probability of finding it in the neighborhood of point x of the domain at a later time t after the deployment? In practice, this is accomplished by identifying some selected regions within the DART area and analyzing the evolution of ensembles of historical drifters launched in these regions. Concentration maps of drifter positions are built as functions of space and time, and they are interpreted in conjunction with other more general statistics, such as the mean flow. The statistics are first performed considering the whole ensemble of historical drifters, while conditional statistics are later introduced depending on the drifters’ seasonal coverage. These concentration maps can be interpreted as PDFs for particle positions, and they can be utilized for several applications. For environmental problems, for instance, they can be used as “risk maps” in the case of pollutant release, indicating the most likely spread of pollutant concentration in space and time. For experimental planning, as in the DART case, the maps can be used to identify the most likely coverage obtained from a given release scheme, therefore providing information on the sampling strategy.

Because in situ datasets are always quite limited compared with numerical datasets, even in highly sampled areas, we expect that the information will be relatively coarse, but it will be able to provide the range of appropriate space and time scales to be considered in most practical applications. Also, we expect that the results will set the ground for further investigations in conjunction with modeling studies, for instance, in terms of validating model results and giving insights into the main factors contributing to uncertainties in the modeled trajectories.

The paper is organized as follows. Section 2 gives a brief background of the Adriatic dataset and the main characteristics of the basin. In section 3, the specific DART area of interest is introduced and the methodology used to analyze the data is described. The results of the analysis are presented in section 4. Conclusions and a brief discussion of the implications for the DART Lagrangian sampling strategy are presented in section 5.

2. The Adriatic Sea and the historical dataset

a. Adriatic morphology and general circulation

The Adriatic Sea is the northernmost semienclosed basin of the Mediterranean Sea, and is connected to the Ionian Sea through the Strait of Otranto at its southern boundary (see Fig. 1 for geographic locations mentioned herein). It can be separated into the following three different subbasins: 1) the southern Adriatic between the Strait of Otranto (to the south) and the Palagruza Sill, with bottom topography as deep as 1200 m in its center, called the South Adriatic Pit (SAP); 2) the middle Adriatic north of the Palagruza Sill, called the Pomo, Jabuka, or Mid-Adriatic Pit (MAP), which contains depressions as deep as 270 m; and 3) the northern Adriatic, a very shallow region (<100 m) to the north characterized by a gentle bathymetry slope. The Adriatic coastline is relatively smooth on the Italian side with the exception of two major capes—the Conero Promontory and the Gargano Cape. In contrast, the eastern coast is rugged and includes hundreds of islands, and the bottom topography is often very steep.

Hydrographic data (Artegiani et al. 1997) and drifter tracks (Poulain 1999, 2001; Ursella et al. 2006, hereafter URS06) show that the mean large-scale, near-surface circulation is cyclonic, with water entering on the eastern flank of the Strait of Otranto and continuing northward along the Albanian and Croatian coasts as a generally weak and wide Eastern Adriatic Current (EAC). The EAC separates into three embedded cyclonic circulation structures by recirculating near the two northern slopes of the SAP and MAP and south of the Istrian Peninsula. All of these branches feed into the Western Adriatic Current (WAC), which extends almost continuously along the Italian coast from the Po River delta in the northern subbasin to the Strait of Otranto. The WAC is swift and concentrated near the coast or minishelf slope and is mainly forced by the buoyancy input from the Po River. Significant seasonal variability of the WAC has been observed with enhanced flow to the southeast in winter and reduced flow in summer. Wind forcing also affects the WAC: during events of strong northeasterly bora winds, the recirculation south of the Istrian Peninsula and the WAC are reinforced; in conditions of southeasterly sirocco winds, the EAC is stronger and the WAC weakens and even reverses (Poulain et al. 2004b).

b. Adriatic historical drifter data

As part of several research and military programs, satellite-tracked surface drifters have been released in the Adriatic Sea since 1990 and have provided data until 1999. Maximum numbers of drifters occurred in 1995 and 1997–98 in concomitance with programs funded by the North Atlantic Treaty Organization (NATO) and the U.S. Office of Naval Research (ONR), respectively (Poulain 2001). The majority of these drifters were released in the southern Adriatic and the Strait of Otranto area. More drifters were deployed in the northern and middle Adriatic between September 2002 and November 2003 as part of the ONR Dynamics of Localized Currents and Eddy Variability in the Adriatic (DOLCEVITA) and NATO Adriatic (ADRIA02/03) projects (URS06).

Most of the Adriatic drifters are Coastal Ocean Dynamics Experiment (CODE) drifters that measure the surface currents within the first meter of water (Davis 1985). Their water-following characteristics have been assessed (P.-M. Poulain and L. Ursella 2006, unpublished manuscript), and it was found that they follow the surface current within 2 cm s−1. All drifters were tracked with the Argos Data Location and Collection System (DCLS) carried by the National Oceanic and Atmospheric Administration (NOAA) polar-orbiting satellites. The drifter position data (latitude and longitude) were edited, interpolated, and low-pass filtered (with a cutoff period of 36 h). The low-pass time series were subsampled every 6 h and the velocity was computed by finite-centered differencing the 6-hourly interpolated/filtered position data. Details on the drifter data processing can be found in Poulain (2001), Poulain et al. (2004a, 2006), and URS06. The drifters used in this study correspond mostly to the years 1992, 1995, 1997–98, and 2002–03.

3. The DART area of interest and the methodology

The DART area of interest is situated in the middle Adriatic, around the Gargano Cape (Fig. 1). It is characterized by a complex topography (see also Fig. 4), with the Palagruza Sill extending from the cape toward the eastern coast of the basin and dividing the two deep depressions of the MAP (to the north) and the SAP (to the south). A well-defined shelf can be seen along the Italian coast, approximately extending up to the 150-m isobath and widening south of the Gargano Cape in the Gulf of Manfredonia.

One of the main goals of the DART experiment is to study the mesoscale instabilities that form in the WAC boundary current near the Gargano Cape and their influence on the downstream circulation. To this end, a suite of different measurements will be used during DART and complemented by a near-real-time modeling effort (see information online at http://oceans.deas.harvard.edu/haley/DART05). An array of current meters will be maintained for extended periods (order of months), while two cruises of 2–3 weeks in winter and summer 2006, respectively, will provide high-resolution synoptic hydrographic data as well as microstructure data. During the first cruise in March 2006, a total of approximately 20 surface drifters will be deployed in the Gargano Cape area. The deployment scheme should provide an approximately homogeneous coverage in space and time during the cruise period. The historical data analysis presented in this paper will contribute to identifying the suitable deployment points, also providing estimates of the expected time scales for drifters to properly cover the area of interest.

As a first step, we subdivide the DART area into appropriate dynamical subregions. The regions are chosen based on the mean flow and on the eddy kinetic energy (EKE) field computed by averaging the whole ensemble of historical drifter velocities over 0.1° square bins (Fig. 2). The mean flow field (Fig. 2a) clearly shows the WAC boundary current flowing southeast along the coast. The current extends approximately up to the 160-m isobath upstream of the Gargano Cape, and has a well-defined core within the 100-m isobath from the coast. Downstream of the cape, the current widens, following the shelf. A small recirculation region can be seen in the lee of the cape, with reduced sampling suggestive of a “shadow zone” resulting from the classical detachment of a boundary current in the presence of an obstacle (Doglioli et al. 2004). The WAC appears connected with the interior flow through two branches of the cross-basin recirculations, along the slopes of the MAP and SAP, respectively. Also, a stagnation point in the mean flow field can be seen on the Palagruza Sill, dividing the MAP and SAP recirculations. In the following we will loosely refer to this point as the “hyperbolic point,” because a hyperbolic (or saddle) point is indeed expected to be present in the specific flow realizations (Poje and Haller 1999). The EKE field (Fig. 2b) indicates that fluctuations are intense in the WAC, especially near the tip of the Gargano Cape and at the southeastern edge of the DART area where the shelf narrows and the current “reattaches” to the coast.

Based on this information, five regions have been chosen and their boundaries are shown in Fig. 2 (thin black lines). Two regions—north polygon (POLYN) and south polygon (POLYS)—are the polygons located inside the WAC upstream of the Gargano Cape. Although POLYN lies almost entirely outside the DART area, it has been included in the analysis because it is representative of the structure of the boundary current before being influenced by the cape, and it can therefore provide an interesting upstream control point. Logistically, ship transects during the cruises are expected to extend up to POLYN, making deployments possible in the region. The other three regions are located downstream of the Gargano Cape: the south Gargano Cape (GARGS) covers the WAC pattern southward of the cape; the north Gargano Cape (GARGN) includes the more complex region close to the hyperbolic point and is influenced by the interior recirculation branch; and the Gargano recirculation (RECIRC) covers the recirculation region in the lee of the cape, in the Gulf of Manfredonia.

For each of these regions, we have selected all the drifters that have either been deployed in the region or have entered it at some time during their life. The deployment and entrance points are considered as initial conditions, that is, of independent trajectory realizations. Because the number of actual deployments is very small with respect to the entrance points, almost all of the ICs are located on the boundaries of the regions (see Figs. 4, 7). Notice that considering the entrance points as independently given initial conditions is an approximation because the entrance points depend on the flow field that advected the drifters. Also, in the practical implementation, a number of drifters cross more than one region in the DART area and their trajectories are therefore included in the analysis of each of the covered regions. Whenever possible, the ICs in each region are further grouped to provide ensembles of dynamically consistent initial conditions. For example, in POLYN and POLYS (see section 4a), the ICs in the core of the boundary current are separated from those in the more external part.

The evolution of particle motion for each IC group is statistically characterized using a number of diagnostics. First, a qualitative assessment is performed by plotting the corresponding “forward” and “backward” trajectories for each region of the DART area. If the drifter IC occurs at a time t0, the forward (backward) trajectories are those corresponding to the time period t0tt0 + tforw (t0tbacktt0), where tforw (tback) is a fixed interval. Clearly, if the IC is an actual deployment, only the forward trajectory is computed. Visual inspection of the trajectory ensembles provides a direct and general assessment of the transport from/to a certain region.

A more quantitative analysis is then performed on the forward trajectories, computing their concentration maps or PDFs over a 0.2° × 0.2° spatial grid at fixed times tforw. In each grid cell, the concentration is computed as the ratio between the number of drifters present in the cell at the time tforw, and the total number of drifters in the considered IC group. The PDF maps are complemented by other bulk descriptors, such as arrival times and ensemble properties. These analyses are first performed over the whole ensemble of drifter trajectories, and then are repeated considering seasonal dependence.

4. Results

A first characterization of the particle evolution is given in Fig. 3 by the forward and backward trajectories (tforw = tback = 10 days) for three significant regions: POLYN in the WAC boundary current upstream of the Gargano Cape, and GARGN and RECIRC downstream of the cape. POLYS and GARGS are qualitatively similar to their corresponding northern counterparts.

As can be seen, particles entering POLYN originate from the northern WAC and from the recirculating cross-basin branch along the MAP. Once they enter the boundary current, drifters tend to stay in it, moving southeastward following the bathymetry. Some escape points are visible, mostly upstream of the cape and at its tip. Particles entering GARGN also originate from two different locations: the upstream WAC and the recirculating SAP branch. After entering GARGN, they are mostly entrained in the boundary current moving southeastward, even though a clear escape point can be seen offshore of the tip of the Gargano Cape, with drifters moving northward. The ensemble of ICs in RECIRC is reduced, because few drifters enter the lee of the cape. They mostly originate from the upstream boundary current and skirt the tip of the cape as they move downstream. They tend to recirculate once or twice in the region and then move southeastward to join the boundary current.

A more quantitative description of the forward motion is given in the following sections, which consider the regions upstream and downstream of the cape separately.

a. Regions upstream of the Gargano Cape

The drifters’ ICs in the two upstream regions, POLYN and POLYS, are shown in the upper panels of Fig. 4. They are further divided in two groups for each region: BCint (filled circles) includes the ICs in the internal part of the WAC, approximately defined as the area within the 100-m isobath from the coast, while BCext (open circles) includes the ICs in the external part of the WAC, between the 100- and 160-m isobaths. The temporal evolution of the PDF for the BCint drifters of POLYN is shown in Fig. 5. The results for POLYS (not shown) are qualitatively similar. In addition to the PDF map corresponding to a particular tforw time value, each panel of Fig. 5 indicates the percentage of drifters that are inside the DART area and the percentage of those that have stopped working at the time tforw. Drifters “die” for various reasons, from battery failure, to beaching, and being caught in fishermen nets. Their “mortality” has to be taken into account because it can significantly alter the number of realizations and even bias the statistics (e.g., Falco et al. 2000). The total number of drifters in BCint is initially equal to 58, while 23 drifters are already dead at tforw = 15 days. This mortality rate is rather high compared with the typical lifetime of a drifter in the Adriatic Sea [mean half-life of 35–40 days, see Poulain (2001) and URS06].

The PDF maps in Fig. 5 show the spreading of the drifters as they move southeastward in the boundary current. The spreading is mostly alongshore and it is likely to be due to the influence of the sheared current, as well as to the effect of having considered many different realizations in computing the statistics. After 2 days, a few drifters have already reached the tip of the Gargano Cape, while the maximum of the concentration is still located at the boundary between POLYN and POLYS. Most of the drifters (70%) are found in the DART area after 8 days, while only 43% are still present after 15 days. While most of them have exited the area following the current to the south, an escape point at the tip of the cape can also be seen.

Drifters with ICs in BCext (33 in total) have a qualitatively similar behavior to those in BCint, but they also show some significant differences (Fig. 6). The alongshore spreading is reduced with respect to BCint, probably because of a weaker current shear. Also, particles show a higher probability of escaping the boundary current to enter the interior, so that a smaller percentage (54%) is found in the DART area after 8 days. An escape point can be seen upstream of the cape, slightly north of the boundary between POLYN and POLYS. The presence of such an escape point is also suggested by the mean flow structure in the corresponding location (Fig. 2a), which shows a (weak) recirculation probably connected to topographic features.

Results for POLYN and POLYS are summarized in Tables 1 and 2. They are obtained by considering the whole ensemble of initial conditions BCint and BCext, and focusing on a time period of 21 days, which approximately corresponds to the duration of the planned DART cruises. The statistics in Table 2 are aimed at answering the following questions: how many drifters reach the tip of the Gargano Cape in 21 days; what is their fate after they pass the tip; and, in particular, how many drifters leave the boundary current to enter the Adriatic interior, and how many stay in the boundary current or recirculate in the lee of the cape?

To answer these questions, two quantities ntot and ndie_earlier are first computed for each region and are listed in Table 1. The quantity ntot indicates the total number of drifters with initial conditions in the region, while ndie_earlier indicates the number of drifters that die during the 21-day period without reaching the tip of the cape. The statistics in Table 2 are computed as percentages with respect to the difference between ntot and ndie_earlier. The ndie_earlier drifters are removed because it is impossible to say whether or not they would have passed the cape had they lived long enough. Specific analyses to address the consequences of mortality and possible associated biases (Falco et al. 2000) are out of the scope of this paper, and will be addressed in future works. The quantities in Table 2 indicate the percentage of drifters that reach the tip of the cape (ngarg), the percentage of those that migrate to the interior after passing the tip (ngarg2int), and the percentage of drifters that enter the Gargano Cape recirculation (ngarg2rec). Also shown is the average arrival time Tm (in days) taken by the ngarg drifters to reach the tip of the cape. When computing these quantities, the tip of the Gargano Cape is identified with the eastern boundary of POLYS. Furthermore, ngarg2int and ngarg2rec are computed by monitoring the trajectories inside the GARGS + GARGN area downstream of the Gargano Cape and by checking whether or not they permanently leave the area during the 21-day period of interest.

The results in Table 2 are qualitatively similar for POLYN and POLYS, indicating that the majority of the trajectories (83%–88%) reach the tip of the cape by remaining inside the boundary current (in agreement with Figs. 5 and 6). After reaching the tip, approximately 10% of the drifters leave the current and penetrate the interior, while the remaining ≈70% keep flowing southeastward in the WAC. Only ≈7% recirculate in the lee of the cape. The arrival times Tm are 9.3 and 5.8 days for POLYN and POLYS, respectively. While these are average times, the distribution of individual arrivals is shown by the histograms in the lower panels of Fig. 4. The distributions appear quite flat, especially for POLYN, in agreement with the alongshore spreading shown in Fig. 5, while for POLYS a small peak can be seen at 2–4 days. Notice that POLYS has nonzero values in the class 0–2 days, mostly resulting from a few ICs that lie very close to or even at the eastern border of POLYS (Fig. 4). As a test, we have removed the ICs in the eastern half of POLYS and we have found differences only in the first histogram class, while the statistics in Table 2 remain basically unchanged.

In summary, the results indicate that drifters with ICs in the WAC upstream of the cape have a very high probability of reaching the cape and staying in the boundary current even after they have passed the tip. This probability is especially high for drifters with ICs in the core of the current, within the 100-m isobath from the coast. A quantitative prediction of the trajectories and of their arrival times at the tip of the cape is, on the other hand, hard to obtain because of the alongshore spreading related to the shear of the mean current and to the high variability of the drifter realizations. We will return to this specific point in section 4c when considering the effects of seasonality.

b. Regions downstream of the Gargano Cape

The drifters’ ICs in the three downstream regions GARGN, GARGS, and RECIRC are shown in Fig. 7. The ICs for GARGN correspond mostly to entrance points from the WAC boundary current, including both BCint and BCext. They are grouped together because the sampling is reduced with respect to POLYN and POLYS (46 drifters in total), and also because they exhibit a similar behavior. The PDFs in Fig. 8 show that the drifter concentration follows two distinct patterns. The main pattern is entrained in the boundary current and moves southeastward, while a secondary concentration branch moves northward toward the Adriatic interior. This is not surprising considering the structure of the mean flow (Fig. 2a) in the northwestern part of GARGN (where the ICs are located), which is an area very close to the hyperbolic point between the two cross-gyre recirculations. The two out-flowing branches of the hyperbolic point correspond to the two main concentration patterns seen in Fig. 8 (see, e.g., the panel for tforw = 4 days). After 8 days, only 34% of the drifters are still within the DART area, with drifters exiting both to the southeast and to the north of the region.

In the GARGS analysis, all the available ICs (104 drifters in total) have been considered because they exhibit a similar behavior. They include drifters entering from the WAC, as well as drifters entering from the cross-basin recirculation around the SAP (Fig. 7). The time evolution of the PDF (Fig. 9) is strikingly different from that for GARGN (Fig. 8), with virtually all of the drifters entrained in the boundary current and moving southeastward. A certain degree of alongshore dispersion can be seen at tforw = 2 days, but it is much weaker than in the upstream boundary current (POLYN; see Fig. 5), probably because the downstream current is wider and less sheared. At the southeastern corner of the DART area, the current narrows and intensifies (Fig. 2a), following the topography. After 8 days, only 28% of the drifters are still in the DART area, and most of the drifters have exited toward the south.

The available ICs in RECIRC (23 drifters in total) are located along the boundary with GARGS (Fig. 7). The PDFs in Fig. 10 show that, while almost all the drifters are trapped in RECIRC during the first 2 days, they start exiting the region after 4 days. Most trajectories loop inside RECIRC once or twice and they then leave the lee of the Gargano Cape. Overall, drifters tend to stay longer in the DART area with respect to drifters for GARGS or GARGN. After 8 days, in fact, 47% of the particles are still in the area. Nevertheless, the RECIRC drifters also eventually tend to be entrained in the boundary current moving southeastward and exiting the area. Notice that the sampling in this region is scarce, suggesting that the lee of the cape tends to be isolated from the main boundary current system for most of the time. The mechanisms that lead to the formation of the recirculation connecting the lee to the main current are not obvious at this time and they will be investigated in future works.

In summary, drifters with ICs in the DART area downstream of the cape have an overall tendency to be entrained in the boundary current. Only in the region offshore of the tip of the cape on the Palagruza Sill is a very high variability observed, with drifters being entrained in a northward branch and moving toward the Adriatic interior. This is related to the presence of a hyperbolic point, which induces significant spatial variability. Also, the exact position of the hyperbolic point is likely to vary in time, so that particle evolution may vary according to the specific realization. In this region, the statistical analysis of historical data can provide only partial answers, and it is expected that the joint use of modeled data will lead to improved results.

c. Seasonal dependence

The results discussed in the previous sections are obtained using the whole historical dataset, and they show a significant variability in the evolution of drifter trajectories. To resolve this variability at least partially, conditional statistics have been considered as a function of seasons. The temporal definition of the seasons used here is similar to the calendar as appears in URS06: spring is from 21 March to 20 June, summer from 21 June to 20 September, fall from 21 September to 20 December, and winter from 21 December to 20 March.

As a first step, the mean flow has been computed for the four seasons. This is displayed in Fig. 11. Although the data coverage is obviously reduced with respect to Fig. 2, and many bins have less than 10 independent measurements, the results are instructive and clearly show the flow evolution. During fall and winter, the WAC is well defined, wide, and energetic. The cross-basin recirculations around the SAP and the MAP are evident, connecting the WAC to the interior and to the EAC. During spring and summer, the WAC appears weaker and more coastal (also resulting from the lack of observations offshore). Recirculation patterns (MAP and SAP) are less evident and the connection of the coastal flow with the Adriatic open sea is reduced. As discussed in a number of previous works (Zore-Armanda 1956, 1969; Artegiani et al. 1997; Hopkins et al. 1999; URS06), the Adriatic circulation and the WAC in particular are strongly influenced by Po River discharges and by wind conditions. The Po River discharge is high in fall and winter with a maximum around December–January (URS06), while it is reduced in summer with a minimum around July and August. This is one of the main reasons for the reduced WAC strength during the summer season. Regarding the wind forcing, strong northeasterly bora winds occurring mostly in fall and winter are associated with reinforced WAC and cross-basin recirculations. During summer, the wind regime is more variable, and it includes episodes of strong southeasterly sirocco winds that tend to weaken the WAC and the cross-basin recirculations (Poulain et al. 2004b).

The seasonal dependence of drifter evolution is studied focusing on the regions upstream of the Gargano Cape because they have an extended sampling and they show a well-defined seasonal signal. The qualitative behavior of drifters in POLYN is illustrated by the backward and forward trajectories in Fig. 12. In fall and winter, consistent with the mean flow pattern (Fig. 11), drifters enter POLYN from the WAC and from the MAP recirculation, while the recirculation branch weakens in spring and almost disappears in summer. Forward trajectories tend to stay in the WAC moving southeastward, although some escape points are evident, especially in the fall, mostly resulting from the northward branch of the recirculation. Summer trajectories are significantly more confined to the coast with respect to the fall and winter periods.

The time evolution of the PDF for the POLYN drifters is shown in Fig. 13. To have sufficient coverage, all of the ICs (BCint + BCext) are considered to compute the drifter concentrations. Furthermore, the maps are drawn for two extended seasons: fall–winter and spring–summer. The sampling is more reduced in spring–summer, with a total of 38 drifters versus 61 in fall–winter. During the first few days, the spring–summer drifters stay close to the coast and show a high alongshore spreading, with some realizations reaching the tip of the Gargano Cape after 2 days and approaching the DART southern border after 4 days. After 8 days, the PDF covers the whole coastal region, with 58% of the drifters in the DART area. Of the remaining drifters, 33% are dead and 9% have left the area to the southeast, following the WAC. In contrast, the fall–winter drifters are less confined to the coast and they have a higher probability of escaping the WAC north and at the tip of the Gargano Cape. After 8 days, the percentage of drifters inside the DART area is approximately the same as that in spring–summer (59%), but the mortality is reduced (27%) and the remaining drifters (14%) have left the area mostly toward the interior. Also, during the first few days, the fall–winter PDFs show a reduced spreading with respect to the spring–summer period, with drifters reaching the tip of the cape after 4 days instead of 2. This result is somewhat surprising considering that the mean flow is generally weaker in the spring–summer period than in fall–winter, but we attribute it to the increased flow variability of the spring–summer seasons that is not resolved by the drifter data coverage. During summer, the WAC is reduced because of the weak Po River discharge, and it is therefore more directly influenced by the wind. Visual inspection of individual trajectories in late spring and summer shows the existence of two distinct regimes. Some trajectories, probably under the effect of strong bora winds, are extremely fast, move down the boundary current, and cover the whole area in a few days. These fast trajectories are responsible for the fast spreading of the spring–summer PDFs. Other trajectories, instead, are trapped in mesoscale structures and stay in the boundary current for much longer. Conditional statistics as a function of wind conditions during summer would help investigate this point further, but unfortunately the sampling is insufficient at this time to conduct such investigation. Another aspect that further complicates the statistical description in summer is that many “slow” drifters die, probably beached at the coast. In particular, 42% of the spring–summer drifters for POLYN die during the first 21 days without reaching the Gargano Cape, versus the 28% for the fall–winter period (Table 3). The spring–summer PDFs are then likely to be biased by the fast summer drifters, and therefore only partially reliable. This aspect will be considered when discussing the bulk statistics in the following.

As in section 4a, for the whole dataset (Table 2) we summarize the seasonal drifter behavior in POLYN and POLYS by considering bulk quantities, which are displayed in Table 4. Because of the reduced sampling, some of the statistics cannot be considered significant. For instance, the percentage of drifters entering the recirculation in the lee of the cape is not reliable because the actual number of drifters ngarg2rec is too small. Furthermore, during the summer, only very few drifters reach the tip of the cape in the first 21 days (ngarg = 4 for POLYN and 8 for POLYS), making it impossible to compute significant statistics. For the other three seasons—fall, winter, and spring—the average arrival time at the cape Tm (computed over more than 10 realizations) is quite similar, ranging between 9–10 days for POLYN and 5–6 days for POLYS. A significant seasonal dependence can be seen, instead, in the percentage of drifters exiting the DART area toward the interior after reaching the Gargano Cape (ngarg2int). During fall and winter, ngarg2int ranges between 10% and 20%, while during spring ngarg2int = 0. This is consistent with the PDFs in Fig. 13.

In summary, particles with ICs in the WAC upstream of the cape have a higher probability of leaving the boundary current toward the interior in fall–winter than in spring–summer. During the summer, when the WAC is narrow and more confined to the coast, the trajectories show a very high variability with two different regimes, probably linked to wind forcing. Episodes of very fast trajectories alternate with episodes in which drifters are slow and trapped in mesoscale coastal structures.

5. Conclusions

In this paper, an analysis of historical drifter data is presented in the area of the DART experiment, in the middle Adriatic around the Gargano Cape. The properties of drifter evolution are investigated as a function of the drifters’ initial conditions, and the specific goal of providing suggestions for the DART Lagrangian sampling strategy is considered. The details of the sampling plan will obviously depend on several factors, such as logistics, ship availability, and interaction with the other experimental and modeling activities. Nevertheless, some general suggestions can be provided through our results. The DART cruise planned for March 2006 will include two successive finescale surveys of 2–3 days, separated by about 4 days. These surveys will mostly cover the POLYS and GARGS regions. One of the goals is to maximize the overlap between the ship-based measurements and the drifter data, both in space and time. In other words, it would be ideal to release drifters so that most of them remain in the survey area while the ship is sampling. If a first portion of the drifters (∼eight units) are deployed upstream along the first transect (in the western part of POLYS), they will typically drift by the tip of the cape after 5–6 days and further stay in the survey area for 5–10 days (Fig. 9). We suggest spreading the deployments across the entire width of the first transect with some drifters inshore of the 100-m isobath and others more offshore. As shown by Figs. 5 and 6, BCint and BCext drifters spread downstream similarly; therefore, we expect that the drifters will cover the DART study area well. The probability that a drifter escapes to the northwest or to the north is relatively negligible and there is no high risk in releasing them along the first transect. We suggest seeding a second group of drifters at the same location along the first transect at the beginning of the second survey. A lag of about 4 days with respect to the first survey, combined with a typical residence time in the DART area of approximately 10 days, should produce a broad sampling of the study region in quasi synopticity with the ship survey. Furthermore, some additional launches can be planned to investigate specific regions, chosen in collaboration with the modeling component of DART. For example, a few drifters (two–three units) could be seeded in the neighborhood of the hyperbolic point in GARGN (Fig. 8) to investigate its synoptic location during the experiment, because it is expected to greatly influence the pattern of transport and detrainment from the boundary current. In addition, one–two drifters could be deployed in the Gulf of Manfredonia to sample its characteristic recirculation, because very few drifters are expected to move into that region (≈7%).

Summarizing, the results indicate that the analysis of historical data can provide very valuable information on statistical particle prediction. A number of applications can be foreseen aside from the discussed experiment planning, which include environmental planning and risk mitigation. The methodology should, however, be further improved in the future in order to make it applicable to more general dynamical problems and in order to address some of the open points identified in this paper.

The method is expected to be particularly well suited for coastal flows with characteristics similar to the ones considered here, that is, with a well-defined current system and a significant topographic control. For flows with higher variability and less constraints, such as open-ocean flows or flows with a strong tidal or mesoscale component related to eddies and waves, it can be envisioned that historical data will have to be augmented by synthetic model trajectories. In addition to providing unlimited Lagrangian datasets for analysis, model studies will also help investigate the existence of flow regimes and the associated behavior of particle spreading. This is expected to be a valuable approach also in our DART area to improve our understanding of periods of enhanced variability, such as the observed summer variability. Because such variability is likely related to wind forcing, which is especially effective during summer when the buoyancy forcing from the Po River discharge is weak, conditional statistics with different wind forcing could be carried out with synthetic trajectories from an appropriate hydrodynamical model. A modeling study component would also help investigate the characteristics of the complex hyperbolic point region on the Palagruza Sill.

Another improvement to the proposed methodology that is possible with broader data coverage is the use of more uniformly distributed initial conditions from which to compute the PDF maps. This could make the PDF statistics more general because they would represent the probability of finding a particle in a certain location given its initial condition in a wide spatial area rather than in a concentrated location. We have chosen our ICs as entrance points to the considered subregions mainly for practical reasons. In fact, our Lagrangian data coverage was not wide enough to provide uniformly spaced initial conditions and, at the same time, ensure the statistical robustness of the PDF maps.

Finally, an aspect to be considered in future investigations is the effect of the drifter mortality on the computed statistics. In our case study, for example, the high mortality observed during the summer is likely to be related to the beaching of the more coastal and slower drifters, which tend to bias the statistics toward the fastest trajectories. We plan to address this issue in a future work using synthetic Lagrangian stochastic model trajectories (R. Barbanti and P.-M. Poulain 2006, unpublished manuscript).

Acknowledgments

The authors are thankful to A. Molcard, T. Ozgokmen, A. Haza, and A. Poje for numerous stimulating discussions and to M. Rixen and E. Coelho for organizing the DART experiment and motivating the work. This research was supported by the Office of Naval Research Grants N00014-05-1-9994 and N00014-03-1-0291 and by the Consiglio Nazionale delle Ricerche (Italy). Thanks to all the individuals who have contributed to the drifter dataset used for this study, and to L. Ursella and R. Barbanti for their help with the drifter data processing.

REFERENCES

  • Aref, H., 1984: Stirring by chaotic advection. J. Fluid Mech., 143 , 121.

  • Artegiani, A., , Bregant D. , , Paschini E. , , Pinardi N. , , Raicich F. , , and Russo A. , 1997: The Adriatic Sea general circulation. Part I: Air–sea interactions and water mass structure. J. Phys. Oceanogr., 27 , 14921514.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davis, R. E., 1985: Drifter observation of coastal currents during CODE. The method and descriptive view. J. Geophys. Res., 90 , 47414755.

  • Doglioli, A. M., , Griffa A. , , and Magaldi M. G. , 2004: Numerical study of a coastal current on a steep slope in presence of a cape: The case of the Promontorio di Portofino. J. Geophys. Res., 109 .C12033, doi:10.1029/2004JC002422.

    • Search Google Scholar
    • Export Citation
  • Dunlap, E., , Tang C. C. L. , , and Wang C. K. , 2004: Comparison of model surface currents and drifter data from the Grand Banks. Canadian Hydrography and Ocean Science Tech. Rep. 236, Department of Fisheries and Oceans Canada, Dartmouth, NS, Canada, 34 pp.

  • Dutkiewicz, S., , Griffa A. , , and Olson D. B. , 1993: Particle diffusion in a meandering jet. J. Geophys. Res., 98 , 1648716500.

  • Falco, P., , Griffa A. , , Poulain P-M. , , and Zambianchi E. , 2000: Transport properties in the Adriatic Sea as deduced from drifter data. J. Phys. Oceanogr., 30 , 20552071.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Griffa, A., , Piterbarg L. I. , , and Ozgokmen T. , 2004: Predictability of Lagrangian particle trajectories: Effects of smoothing of the underlying Eulerian flow. J. Mar. Res., 62 , 135.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hopkins, T. S., , Artegiani A. , , Kinder C. , , and Pariante R. , 1999: A discussion of the northern Adriatic circulation and flushing as determined from the ELNA hydrography. The Adriatic Sea, Ecosystem Rep. EUR 18834, Vol. 32, European Commission, 85–106.

  • Kuznetsov, L., , Toner M. , , Kirwan A. D. Jr., , Jones C. K. R. T. , , Kantha L. H. , , and Choi J. , 2002: The Loop Current and adjacent rings delineated by Lagrangian analysis of the near-surface flow. J. Mar. Res., 60 , 405429.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Paris, C. B., , Cowen R. K. , , Claro R. , , and Lindeman K. C. , 2005: Larval transport pathways from Cuban snapper (Lutjanidae) spawning aggregations based on biophysical modeling. Mar. Ecol. Prog. Ser., 296 , 93106.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poje, A. C., , and Haller G. , 1999: Geometry of cross-stream mixing in a double-gyre ocean model. J. Phys. Oceanogr., 29 , 16491665.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., 1999: Drifter observations of surface circulation in the Adriatic Sea between December 1994 and March 1996. J. Mar. Syst., 20 , 231253.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., 2001: Adriatic Sea surface circulation as derived from drifter data between 1990 and 1999. J. Mar. Syst., 29 , 332.

  • Poulain, P-M., , Barbanti R. , , Cecco R. , , Fayos C. , , Mauri E. , , Ursella L. , , and Zanasca P. , cited. 2004a: Mediterranean surface drifter database: 2 June 1986 to 11 November 1999. OGS Tech. Rep. 78/2004/OGA/31. [Available online at http://poseidon.ogs.trieste.it/drifter/database_med.].

  • Poulain, P-M., , Mauri E. , , and Ursella L. , 2004b: Unusual upwelling event and current reversal off the Italian Adriatic coast in summer 2003. Geophys. Res. Lett., 31 .L05303, doi:10.1029/2003GL019121.

    • Search Google Scholar
    • Export Citation
  • Poulain, P-M., , Barbanti R. , , Ursella L. , , and Cecco R. , cited. 2006: Adriatic surface drifter database: 21 September 2002 to 29 February 2004. OGS Tech. Rep. [Available online at http://poseidon.ogs.trieste.it/doga/sire/dolcevita/database_ dolcevita/.].

  • Samelson, R. M., 1996: Chaotic transport by mesoscale motions. Stochastic Modeling in Physical Oceanography, R. J. Aller, P. Miller, and B. Rozovskii, Eds., Birkhauser, 423–438.

    • Search Google Scholar
    • Export Citation
  • Spaulding, M. L., , Isaji T. , , Allen A. , , Hall P. , , and Howelett E. , 2005: Application of a random flight model in search and rescue model systems. Third Lagrangian Analysis and Prediction of Coastal and Ocean Dynamics (LAPCOD) Meeting, Lerici, Italy, ONR and Cosponsors. [Available online at http://www.rsmas.miami.edu/LAPCOD/2005-Lerici/.].

  • Thompson, K. R., , Sheng J. , , Smith P. C. , , and Cong I. , 2003: Prediction of surface currents and drifter trajectories on the inner Scotian Shelf. J. Geophys. Res., 108 .3287, doi:10.1029/2001JC001119.

    • Search Google Scholar
    • Export Citation
  • Thomson, D. J., 1987: Criteria for the selection of stochastic models of particle trajectories in turbulent flows. J. Fluid Mech., 210 , 113153.

    • Search Google Scholar
    • Export Citation
  • Ursella, L., , Poulain P-M. , , and Signell R. , 2006: Surface drifter derived circulation in the northern and middle Adriatic Sea: Response to wind regime and season. J. Geophys. Res., 111 .C03S04, doi:10.1029/2005JC003177.

    • Search Google Scholar
    • Export Citation
  • Wiggins, S., 1992: Chaotic Transport in Dynamical Systems. Springer, 301 pp.

  • Zore-Armanda, M., 1956: On gradient currents in the Adriatic Sea. Acta Adriat., 8 , 6. 138.

  • Zore-Armanda, M., 1969: Water exchange between the Adriatic and the Eastern Mediterranean. Deep-Sea Res., 16 , 171178.

Fig. 1.
Fig. 1.

Description of geographic locations in the Adriatic Sea. The topography of the area is also shown (with 50-m contour levels).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 2.
Fig. 2.

(top) Mean flow obtained by averaging the historical drifter velocities over 0.1° square bins. Gray vectors are for bins with more than 5 but less than 10 independent measurements, n* (where n* is computed considering a Lagrangian decorrelation time scale TL = 2 days); black vectors are for bins with more than 10 independent measurements. The standard error ellipses are computed with respect to the major and minor axis of variability. Also shown are the 100- and 160-m isobaths. (bottom) EKE field obtained by averaging the Lagrangian eddy velocities over 0.1° square bins. Only results for bins with n* ≥ 5 are shown. The thick black lines identify the DART area of interest, while the thin lines identify the following 5 subregions: POLYN (north polygon), POLYS (south polygon), GARGN (north Gargano Cape), GARGS (south Gargano Cape), and RECIRC (Gargano recirculation).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 3.
Fig. 3.

Forward and backward trajectories of the drifters from/to (top) POLYN, (middle) GARGN, and (bottom) RECIRC; the corresponding subregions are colored in gray for clarity. Both tforw and tback are equal to 10 days.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 4.
Fig. 4.

(top) Initial positions of the POLYN and POLYS drifters. Filled (open) circles identify the BCint (BCext) drifters. (bottom) Histograms of the arrival time taken by the ngarg drifters (see Table 2) to reach the Gargano Cape from POLYN and POLYS.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 5.
Fig. 5.

Maps of drifter concentration computed over a 0.2° × 0.2° spatial grid at various fixed times tforw for the BCint drifters from the POLYN region. The initial number of these drifters is 58. See text for percentage definitions.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 6.
Fig. 6.

As in Fig. 5, but relative to the BCext drifters from POLYN (their initial total number is 33).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 7.
Fig. 7.

Initial positions of the BCint and BCext drifters in (top) GARGN, and initial positions of the total drifters in (middle) GARGS and (bottom) RECIRC.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 8.
Fig. 8.

As in Fig. 5, but relative to the sum of the BCint and BCext drifters from GARGN (their initial total number is 46).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 9.
Fig. 9.

As in Fig. 5, but relative to the total drifters from GARGS (their initial number is 104).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 10.
Fig. 10.

As in Fig. 5, but relative to the total drifters from RECIRC (their initial number is 23).

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 11.
Fig. 11.

Mean flow obtained by averaging the drifters velocities over 0.1° square bins for the four seasonal periods. Gray vectors are for bins with more than 3 but less than 10 independent measurements; black vectors are for bins with more than 10 independent measurements.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 12.
Fig. 12.

(left) Forward and (right) backward trajectories relative to the POLYN subregion, during the four different seasons. The time intervals tforw and tback are equal to 10 days.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Fig. 13.
Fig. 13.

As in Fig. 5, but relative to the total drifters from POLYN (Bcint + BCext), during two extended seasonal periods: the (left) fall–winter and (right) spring–summer periods.

Citation: Journal of Atmospheric and Oceanic Technology 24, 2; 10.1175/JTECH1969.1

Table 1.

Total number of drifters with ICs in POLYN and POLYS (ntot), and number of drifters that die during a 21-day period without reaching the tip of the Gargano Cape (ndie_earlier).

Table 1.
Table 2.

Statistics obtained from the total drifters (BCint + BCext) for POLYN and POLYS over a period of 21 days. The percentages are computed with respect to the difference between ntot and ndie_earlier previously displayed in Table 1. See text for definition of the displayed quantities.

Table 2.
Table 3.

As in Table 1, but relative to the four seasons. Notice that a percentage of the ndie_earlier drifters are not truly dead, but have just switched from one season to the other.

Table 3.
Table 4.

As in Table 2, but relative to the four seasons.

Table 4.
Save