Start Submission

Reading: The influence of mesoscale and submesoscale circulation on sinking particles in the northern...


A- A+
Alt. Display

Research Article

The influence of mesoscale and submesoscale circulation on sinking particles in the northern Gulf of Mexico


Guangpeng Liu ,

School of Earth and Atmospherics Sciences, Georgia Institute of Technology, Atlanta, Georgia, US
X close

Annalisa Bracco,

School of Earth and Atmospherics Sciences, Georgia Institute of Technology, Atlanta, Georgia, GE
X close

Uta Passow

Marine Science Institute, University of California Santa Barbara, California, US
X close


Mesoscale eddies and fronts in the ocean greatly impact lateral transport and in turn the trajectories of sinking particles. Such influence was explored for April and October 2012 in the Gulf of Mexico using numerical simulations performed with a regional model at 1-km horizontal resolution. Results are compared qualitatively to field samples from two sediment traps located at GC600 (27°22.5 N, 90°30.7 W) and AT357 (27°31.5 N, 89°42.6 W), 81 km apart. In April the traps collected a comparable amount of material, while in October the flux at GC600 greatly exceeded that at AT357. Through inverse calculations, several thousand particle trajectories were reconstructed multiple times from the ocean surface to the depth of the traps (approximately 1,000 m) using a range of sinking velocities, 20–100 m d–1. Taken together, model results and trap data indicate that cross-shore transport of riverine input induced by mesoscale eddies, and convergence and divergence processes at the scale of a few kilometers, significantly impact the trajectory of sinking particles. The large majority of modeled particles reach the bottom faster than would be expected by their sinking speeds alone. This finding is associated with submesoscale-induced horizontal convergence in the mixed layer that aggregates particles preferentially in downwelling regions, accelerating their descent. Furthermore, this study confirms that the cone of influence of vertical fluxes is highly variable in both space and time in the presence of an energetic eddy field, especially for particles with sinking velocity of 50 m d–1 or less. It also demonstrates that the variability of vertical fluxes in the Gulf of Mexico is highly complex and can be understood only by considering the mesoscale circulation and seasonal cycle of primary productivity, which in turn are linked to riverine inputs, wind forcing and the seasonal cycle of the mixed-layer depth.

Knowledge Domain: Ocean Science
How to Cite: Liu, G., Bracco, A. and Passow, U., 2018. The influence of mesoscale and submesoscale circulation on sinking particles in the northern Gulf of Mexico. Elem Sci Anth, 6(1), p.36. DOI:
 Published on 23 Apr 2018
 Accepted on 27 Mar 2018            Submitted on 06 Jul 2017
Domain Editor-in-Chief: Jody W. Deming; Department of Biological Oceanography, University of Washington, US
Associate Editor: Laurenz Thomsen; Department of Earth and Space Sciences, Jacobs University Bremen, DE


Gravitational settling of particles from the euphotic zone to the deep ocean plays an important role in controlling the ocean’s carbon cycle (Armstrong et al., 2001; Son et al., 2009; Alonso-González et al., 2010; Kiko et al., 2017). Particles sinking rapidly enough to reach significant depths before being degraded or solubilized also contribute to the biological pump (Buesseler et al., 2007; Passow and Carlson, 2012). These particles consist of various organic and inorganic materials such as plankton, detritus, bacteria and minerals (Alldredge and Silver, 1988; Simon et al., 2002; Passow et al., 2012). They are commonly defined as composite particles with size greater than 0.5 mm and include predominately large fecal pellets and marine snow, which is very sticky and may collect oil, pollutants and other substances from the water column (Passow et al., 2012).

Deep fluxes of sinking particles are measured using time-series sediment traps, and fluxes are often interpreted as originating from a Gaussian-shaped catchment area, a few tens of kilometers large, located above the trap (Asper, 1987; Waniek et al., 2000). As pointed out by Siegel et al. (1990), however, horizontal advection and mesoscale circulations need to be accounted in the interpretation of sediment trap data. Through a numerical analysis of Lagrangian particles with different sizes and sinking velocities, Siegel et al. (1990) evaluated the variability in the catchment area as a function of sinking velocity and presence/absence of mesoscale eddies, and found that the catchment area may increase by several tens to hundreds of kilometers if mesoscale eddies are present. Those results were confirmed by observational (Neuer et al., 1997; Guidi and Stemmann, 2007) and numerical works (Siegel and Deuser, 1997; Waniek et al., 2000) and suggest that under certain conditions horizontal advective motions can transport sinking particles laterally from afar. In other words, the particles found in the ocean mixed-layer above a deep trap may not be represented in the material collected by the trap.

Several studies have shown that in and near the mixed layer both the horizontal distribution and vertical mixing of biological tracers, such as biomass, nutrients, and particulate organic carbon (POC), depend largely on mesoscale geostrophic eddies (Martin and Richards, 2001; Martin et al., 2003; McGillicuddy et al., 2007; Lévy et al., 2009; Chelton et al., 2011; Lehahn et al., 2011). The residence time of particulate matter in the surface ocean may indeed be of the same order of magnitude as that of mesoscale motions (several days to weeks; Mahadevan, 2016), providing enough time for biological processes such as uptake of nutrients and transformation of sinking or suspended particles to occur (Lévy et al., 2012a).

Variability at even smaller scales may also play a role (Lévy et al., 2012b; Zhong and Bracco, 2013; Mahadevan, 2016; Zhong et al., 2017). Submesoscale circulations impact lateral mixing on spatial scales from 100 m to 10 km. They are broadly characterized by a Richardson (Ri) and Rossby (Ro) number of order 1 and by vertical velocities comparable to those found in frontal systems (10–100 m d–1; Thomas and Ferrari, 2008). Ri, defined as g’∇ρ/ρ(∇u)2, where g’ is reduced gravity, ρ is density and u is the flow velocity, quantifies the relative role of buoyancy and flow shear in the flow dynamics; Ro, defined as u/Lf, where L is the length scale of the circulation considered and f the Coriolis parameter, estimates the relative role of inertial and Coriolis forces in the flow dynamics. Submesoscale circulations can also influence plankton dynamics, being on the timescale of hours to days comparable to that of phytoplankton growth. For example, it has been shown that submesoscale eddies have been shown to contribute to the onset of plankton blooms by restratifying the mixed layer in early spring (Mahadevan et al., 2012), submesoscale filaments around the eddy perimeter subduct large amounts of POC during spring blooms at high latitudes (Omand et al., 2015), and submesoscale fronts can converge effectively and quickly any floating material into ‘lines’ a few meters in width and up to several kilometers in length (Zhong et al., 2012; Mahadevan, 2016). Submesoscale fronts are particularly abundant in the northern Gulf of Mexico (GoM) where surface density gradients are fueled by the energetic mesoscale field and also by the abundant riverine input (Zhong et al., 2012; Luo et al., 2016; Barkan et al., 2017a).

Our quantitative understanding of the role of mesoscale and submesoscale dynamics in regulating the transport pathways of sinking particles in the GoM remains incomplete. In this work we focus on the northern part of the basin where the inventory of particles that sink from the euphotic layer to the bottom is mostly controlled by the large riverine input from the Mississippi-Atchafalaya River system (Lohrenz et al., 1997; Dagg and Breed, 2003; Giering et al., 2018), and by the presence of over 2,000 natural hydrocarbon seeps (MacDonald et al., 2002). Most of the nutrients that fuel biogeochemical interactions near the ocean surface, and in turn the production of particulate organic matter through photosynthesis and decomposition processes (Lohrenz et al., 1997; Dagg and Breed, 2003), have riverine origin. The Mississippi-Atchafalaya River system contributes 90% of the total nitrogen and 87% of the total phosphorous load to the Gulf (Dunn, 1996; Cardona et al., 2016). A smaller, but not negligible, contribution to the basin’s nutrient budget is also given by oils and gases from natural seeps that, rising to the surface, transport nutrients from deep, nutrient-rich waters upwards into the euphotic layer through an ‘added mass’ bubble effect (D’souza et al., 2016). The transport and spreading of these nutrients within the euphotic layer is then facilitated by the mesoscale (10–250 km scale) circulations. In the GoM they consist of the Loop Current (LC), its detached mesoscale anticyclonic (rotating in the clockwise direction) Loop Eddies or Rings that are surrounded by strong vorticity filaments of opposite sign (Zhong and Bracco, 2013; Zhong et al., 2017), and in a large number of smaller cyclonic and anticyclonic eddies (Luo et al., 2016). Biggs and Müller-Karger (1994), for example, showed that cyclonic and anticyclonic eddies shed by the LC can transport nearshore water seaward at least 100–200 km, while Del Castillo et al. (2001) and Gilbes et al. (2002) provided evidence of intrusion events of the Mississippi River plume offshore using in situ measurements of riverine materials (i.e., detrital colored organic matter and chlorophyll a [Chl a]) and ocean color data. Numerical studies further support a role for mesoscale structures in the cross-shelf transport of river-discharged water. For example, Toner et al. (2003) applied a Lagrangian analysis to satellite-derived Chl-a plumes to show that the inter-eddy advection transports organic material from the shelf to the ocean interior, and Morey et al. (2003b) noticed a remarkably high-Chl-a, low-salinity tongue extending from De Soto Canyon toward the southeast along the eastern edge of the LC by analyzing 14 years of model simulation. Overall, these works argue for linkages between the LC, the mesoscale eddies and the riverine freshwater input that contributes to the spatial and temporal variability of the formation of particulate matters. They focus, however, on the upper ocean, and little is known about the interactions between mesoscale advection and vertical transport of particulate matter through the whole water column.

Here we have used integrations from a submesoscale-permitting three-dimensional regional ocean model to test the hypothesis that mesoscale and submesoscale variability can at times significantly impact the catchment area of sediment traps in the northern GoM. We considered specifically two traps deployed 81 km apart, each at a depth greater than 1000 m in the northern GoM and 120 m above the bottom, as part of the ECOGIG (ECOsystem impacts of Oil & Gas Inputs to the Gulf; project. Through the analysis of Lagrangian backtracked trajectories, we explored the temporal and spatial variability of the sinking particles, providing an interpretation of the variability observed in the trap data over two collection periods, in April and in October 2012. We focused on these periods because the traps collected comparable amounts of POC in April yet very different amounts in October, leading us to hypothesize that mesoscale circulation contributed to the differences between the periods.

Our goal was to understand if and how much the physical circulation and in particular Loop Current eddies affect the catchment area of sinking particles. As we did not aim at reproducing the observations in the most realistic way, we have not used a data-assimilative model. We also integrated trajectories of passive Lagrangian sinking particles with vertical velocities between 20 m d–1 and 100 m d–1 without accounting for aggregation or dissolution, imposing constant size-related sinking velocities throughout the water column.

Materials and Methods

Sediment traps, Chl-a data and riverine discharge

The traps considered in this work are at the Green Canyon lease block GC600 (27°22.5’ N, 90°30.7’ W, 1380 m deep; trap depth 1260 m), the largest natural hydrocarbon seep in the northern GoM (Roberts et al., 2010), and at a nearby non-seep reference site in the Atwater Valley lease block AT357 (27°31.5’ N, 89°42.6’ W, 1160 m deep; trap depth 1040 m). The seepage activity at GC600 has been linked to a localized increase in primary productivity and surface Chl-a concentrations near the ocean surface (D’souza et al., 2016).

Both traps collected settling particles in 11–28-day intervals from April 2012 onward. In this work we concentrated on two collection periods, 20–30 April and 2–12 October 2012. A detailed analysis of the sediment trap data from spring of 2012 to spring of 2016 at GC600 and for summer of 2015 for AT357, with an attempt to attribute their variability to riverine outflow, seasonality of primary productivity and presence/absence of natural seeps can be found in Giering et al. (2018). While we focused on the time series of POC, most flux compounds measured were significantly correlated, and in particular the total mass flux with POC, and the particulate organic nitrogen (PON) with POC.

Additionally, daily Chl-a data for 20–30 March and 1–15 September 2012 at 4-km resolution were downloaded from NASA’s Goddard Space Flight Center ( The chosen periods represent the time at which the particles collected in the traps where likely near the ocean surface. Sea surface height (SSH) anomaly products used for model validation are calculated based on AVISO altimetry data and were obtained from the Gulf of Mexico Coastal Ocean Observing System (GCOOS, Finally, data on freshwater discharge from the Mississippi-Atchafalaya system were obtained from the USGS ( at daily frequency and averaged to match the trap collection periods.

Model configuration, validation and particle deployment

We used three-dimensional velocity fields generated by a regional ocean model, the Regional Ocean Modeling System (ROMS) (Marchesiello et al., 2003), to backtrack the trajectory of sinking particles from a depth of about 1000 m, at the grid points where the two ECOGIG sediment traps were located (Figure 1), to the ocean surface. Thus we can identify the specific source areas of the particles reaching the traps.

Figure 1 

Bathymetry of the Gulf of Mexico. The topographic data are derived from the 2-min Gridded Global Relief Data Collection topography ETOPO2. Sediment traps were set 120 m above the seafloor at GC600 (black plus symbol) and AT357 (green plus symbol). Major topographic features are also indicated. DOI:

ROMS is a three-dimensional, free-surface, hydrostatic ocean model that we used to simulate the circulation in the GoM with a resolution of 1 km in the horizontal to partially resolve submesoscale dynamics and 50 stretched, terrain-following layers in the vertical. Biharmonic horizontal diffusion and the Mellor-Yamada vertical mixing scheme with level 2.5 closure (Mellor and Yamada, 1982) were applied as mixing parameterizations. The model bathymetry is derived from the 2-min Gridded Global Relief Data (ETOPO2) topography (Sandwell and Smith, 1997) interpolated to the model grid and modified to reduce horizontal pressure gradient errors using the Sikiric et al. (2009) method, with maximum slope factor (rx0) of 0.25 and maximum hydrostatic inconsistency number (rx1) of 15. The model domain extends over the whole GoM, covering the region between about 98°–82° W and 22°–32° N (Figure 1). It has open boundaries to the east, north and south sides that were nudged to the 6-hourly Hybrid Coordinate Ocean Model – Navy Coupled Ocean Data Assimilation (HYCOM-NCODA) Analysis system (GOMI0.04/expt_31.0, Six-hourly wind stresses and heat fluxes as well as daily precipitation from the European Centre for Medium-Range Weather Forecast ERA-interim reanalysis were used to force the model (Poli et al., 2010). More details about this configuration can be found in Bracco et al. (2018).

For each trap collection period (20–30 April and 2–12 October 2012) we performed a 65 day-long integration covering the two months prior to the collection, each initialized from the HYCOM-NCODA hindcast. The first two weeks in each run were discarded as adjustment/spin-up time. The seasonality and impact of the freshwater fluxes from the Mississippi-Atchafalaya River system on the sea surface salinity (SSS) fields were retained in the initial conditions of each run with no further river outflow implementation in the fall case, while nudging to the surface salinity field from the HYCOM-NCODA hindcast was applied at the end of the spin-up in the spring case. In early spring the riverine outflow increased significantly between 18 February 2012, when our simulation was initialized, and early-to-mid March, when particles collected in the trap were assumed to be at the surface. The subsequent nudging to HYCOM-NCODA guarantees proper simulation of the generation of submesoscale circulations fueled by the lateral density gradients induced by the riverine discharge propagating into the Gulf (Luo et al., 2016). In fall of 2012, on the other hand, the influx of river water was small in August and early September, and no further nudging was required for a proper representation of the salinity distribution in the region of interest. The modeled SSS compares very well to the HYCOM-NCODA hindcast in the first three weeks of both simulations, when the particles were near the surface, but is underestimated by the end of them, mostly over the shelves. In the region of interest of this work, i.e., offshore the northern shelf of the GoM, the differences between the spatially averaged SSS anomalies in the hindcast and in our ROMS runs remain less than 8% by the end of the third week.

Neglecting the momentum flux associated with the river outflow makes this configuration unsuitable for exploring the origin of particles collected by traps close to the Mississippi River mouth and directly influenced by the river ‘jet’ (Barkan et al., 2017b), as in the case of data from OC26 (88.36° W, 28.70° N; Giering et al., 2018), but provides realistic surface salinities and salinity gradients for locations 100 km or more from the river mouths, as in the case of GC600 and AT357.

ROMS uses initial and boundary conditions from a hindcast that assimilates all available satellite and in-situ observations, but is run without any data assimilation. The mesoscale circulation can therefore deviate from that of HYCOM and from observations, and can do so significantly over the Eulerian time scale of the flow, which was estimated to be about 50 days in the GoM (Cardona and Bracco, 2016). Figure 2 presents a qualitative validation of the modeled mesoscale circulation over two days during which some of the particles collected at the trap locations in the relevant time window were backtracked by ROMS to the ocean surface. It compares SSH maps from the model with those from multiple altimetry satellites compiled on a 0.25° × 0.25° grid by the Colorado Center for Astrodynamics Research. Although the model slightly overestimates the magnitude of SSH in both seasons, the spatial pattern and in particular the location of major eddies and of the LC are well represented. The overestimation may be due to the stress formulation of the wind forcing that does not account for eddy-current feedbacks (Renault et al., 2016). We note, however, that small mismatches between modeled and observed circulation may translate into large differences in particle transport, and that our results are intended to shed light on processes that can impact the cone of influence of sediment traps in the Gulf, without hindcasting or forecasting aspiration.

Figure 2 

Validation of modeled sea surface height and modeled salinity fields. Sea surface height patterns based on AVISO satellite data downloaded from the Gulf of Mexico Coastal Ocean Observing System (GCOOS, panels a and b) and on model data (panels c and d), as well as modeled salinity field in the GoM (panels e and f). GC600 and AT357 sites are indicated by black and green plus symbols, respectively. Major surface circulation features such as the Loop Current and Loop Current anticyclonic eddy are shown in panels a and b. Arrows in panels c and d represent current vectors. DOI:

For the objective of this work, validating the modeled vertical transport would also be important. Such validation, however, is complicated by the paucity of observations of vertical velocity, w, in the ocean. A 16 month-long record of w was collected in the GoM at one location, through a mooring equipped with two upward-looking 75 kHz acoustic Doppler current profilers (ADCPs), between 2003 and 2004 (Rivas et al., 2008). An investigation of w variability, in a nearly identical model configuration to that adopted here and with a comparison to the ADCP data, is contained in Zhong et al. (2013), indicating that ROMS captures the observed w intensity and variability in association with the LC eddies.

Finally, we verified that the differences in topographic slopes at the two trap locations are small, at least in the model, and that lateral velocities at the depth of the traps (120 m above the bottom) are comparable (Figures S1 and S2) during both of the seasons considered. These verifications ensure that the lateral flux of particles into the traps is not strongly influenced by bathymetric features.

In the following we describe the evolution of 3698 Lagrangian tracers backtracked multiple times from the trap locations to the ocean surface off-line using hourly averaged three-dimensional velocity data generated by ROMS. The error in the tracer trajectories introduced by off-line tracking is negligible when using hourly flow fields at the resolution considered (Bracco et al., 2016). The tracers were uniformly deployed at about 100 m above seafloor (about 1250 m at GC600 and 1050 m at AT357) and their trajectories were integrated using LTRANS v.2b (Schlag and North, 2012). Most of the organic aggregates collected in sediment traps are commonly regarded as marine snow (Siegel et al., 1990). The sinking velocity of marine snow varies widely, depending on size, porosity and component particles, ranging between a few meters to several hundred meters per day (Asper, 1987; Alonso-González et al., 2010; Passow et al., 2012). While sinking through the water column, the flux of marine snow declines with depth, due to losses to remineralization and grazing (De La Rocha and Passow, 2007; Buesseler et al., 2008). Slowly sinking aggregates (sinking speed < 10 m d–1) have a very high chance of not arriving at the bottom. Large particles with rapid sinking velocities (e.g., > 150 m d–1) have a short residence time in the water column and are not impacted or very moderately impacted by lateral advection. We therefore neglected both very slow- and very fast-sinking particles and explored the 3-dimensional trajectories of passive particles with sinking velocities of 20 m d–1, 30 m d–1, 50 m d–1 and 100 m d–1 which were taken to be constant through the water column. For each case, we considered an ensemble of six (for 20 m d–1) or seven integrations performed by releasing particles once a day over a week.

Results and Discussion

Trap data in April and October 2012

In this work we focused on the April and October sediment trap collections to investigate the hypothesis that the mesoscale circulation may explain similarities and differences between these two periods. Figure 3 presents the POC concentrations measured by the two traps in 2012. The concentrations were similar in the first two months, diverged slightly from mid-June to early July, were comparable again in the second half of July and August, and presented large differences in fall and through the end of the year, with the largest discrepancies recorded in early October.

Figure 3 

Time series of particulate organic carbon (POC) data collected from the sediment traps. The sediment traps were deployed at 120 m above the seafloor at GC600 (black) and AT357 (green); data shown cover the period from initial deployment in April 2012 to the end of that year. Pink-shaded boxes highlight the two periods of interest: 20–30 April when the magnitudes of POC flux were similar; and 2–12 October when much higher concentrations of POC were found at GC600 than at AT357. DOI:

POC concentrations generally co-vary with Chl a and, in the GoM, with the input of organic material in the basin through riverine discharge. Although directly impacted by nutrient inputs through the riverine system, Chl-a concentrations also depend on other physical conditions. In the northwestern portion of the Gulf, where the traps were located, space-averaged Chl-a anomalies have been observed to increase after episodic wind events, which are frequent in winter and have a climatological seasonal cycle that generally follows that of the mixed-layer depth. As a result, Chl-a concentrations decrease from a winter maximum, generally in February to early March when the mixed layer is deep, to a minimum that can vary significantly from year to year but is most commonly recorded in late August and September, when the mixed-layer is shallowest (Muller-Karger et al., 2015). River discharge, on the other hand, is usually greater in spring and reaches its minimum in fall, with large interannual changes (Cardona et al., 2016).

For the purpose of this work we compared the time series of freshwater discharge from the Mississippi-Atchafalaya system with the POC time series from April 2012 to August 2014 and during February 2015 at AT375 and GC600, respectively (trap records were interrupted for a few months after these dates; Giering et al., 2018), as shown in Figure S3. Despite the small distance between the traps, the influence of the discharge was different, with POC concentrations at AT375 co-varying with the outflow (maximum correlation with the discharge two months prior to collection; r = 0.61, statistically significant at the 99% level), and no correlation evident at GC600.

Mesoscale and submesoscale circulations in March/April and September/October 2012

The mixing and stirring processes induced by mesoscale and submesoscale activities play a central role in controlling the dispersion of biological and physical tracers in the ocean. The mesoscale circulation in the GoM is dominated by the presence of the LC and the LC eddies. The LC enters the GoM basin through the Yucatan Channel and leaves through the Straits of Florida (Weisberg et al., 2000; Alvera-Azcárate et al., 2009). It influences the south part of the domain directly and is usually confined to the east of 86°W (Vukovich, 2007). The northern boundary of the LC penetrates northward to about 26.5–27.5°N latitude based on previous observational and numerical studies of the LC intrusion (Vukovich, 1988, 2007). At irregular intervals the LC then retreats southward after the detachment of a large LC eddy. On average, eddy shedding takes place every 11 months, but the variance characterizing the shedding time is very large (Sturges and Leben, 2000). These eddies, with diameters of several hundred kilometers, move westward across the GoM at speeds of about 5 km day–1 (Elliott, 1982; Vukovich, 2007) and finally decay against the continental margin or merge with eddies existing in the western GoM. For this work, we note that AT357 is located at the western boundary of the Mississippi Slope (Figure 1). This topographic feature is sufficiently large and tall to interfere with the propagation of the LC and its detached eddies, which can be found more often in the proximity of GC600 but rarely atop AT375.

The GoM mesoscale circulation patterns differed appreciably between the two periods considered in this work. Going back to March, two anticyclonic features, a large LC eddy west of 90°W and the LC extension reaching as far as 27°N, occupied the basin (Figure 2), but no large mesoscale structure was found directly above the traps. In September, on the other hand, the central portion of the basin was dominated by a large anticyclonic LC eddy centered at about 25.5°N, 90°W that resulted from the shedding of the LC extension seen in March and extended as far as GC600, while the LC extended only to the southeastern portion of the Gulf.

The mesoscale variability is reflected in the salinity maps, with the LC water of tropical Atlantic origin carrying a fresher signature than the surrounding water. Significant differences between March and September 2012 can be seen also in the horizontal distributions of SSS. Low-salinity waters were mostly trapped in the coastal region in March, but spread more widely offshore in September. This difference is caused by the seasonal variation of wind stress in the Gulf (Morey et al., 2003). The freshwater discharged from the Mississippi-Atchafalaya River system moved toward the west in spring, along the broad, shallow Texas-Louisiana shelf, and then southward, remaining more effectively confined along the Mexico coastline. In summer and fall, on the other hand, the freshwater masses moved predominately eastward, encroaching more easily on the LC and the numerous eddies that populated the offshore waters. Those mesoscale circulations then carried them offshore through mixing and stirring, and in the year of this study towards the location of the traps.

At scales of a few kilometers, mesoscale eddies smaller than the LC ones, both cyclonic and anticyclonic, and vorticity filaments can be identified in both months offshore the shelf in the vorticity field shown in Figure 4. Relative vorticity, defined as ζ = ∂v/∂x – ∂u/∂y, where u is velocity along the x (longitudinal) direction, and v is velocity along the y (meridional) axis, quantifies the spinning (clockwise if negative or anticlockwise if positive) of the water parcels. In Figure 4, relative vorticity is plotted normalized by the Coriolis parameter f, providing a quantification of the local Rossby number (Ro = |ζ|/f). Ro of ~1 or greater is indicative of ageostrophic motions; i.e., of circulations for which the balance between the pressure gradient force and the Coriolis force is not attained, as commonly found in the case of submesoscale dynamics. In between mesoscale structures, a large number of submesoscale circulations such as vorticity filaments and small vortices, on a spatial scale of 1–10 km and with a lifespan of days, can be seen, particularly in September. Submesoscale processes extract available potential energy from the mixed layer (Molemaker et al., 2005; McWilliams, 2008) through processes such as frontogenesis and mixed layer instabilities (Thomas and Ferrari, 2008; Luo et al., 2016; Barkan et al., 2017a). Submesoscale features are stronger and more numerous when more available potential energy can be extracted. In general, for open ocean conditions, these characteristics imply a seasonal cycle of submesoscale activity that follows that of the mixed layer depth (Capet et al., 2008; Mensa et al., 2013, Qiu et al., 2014; Callies et al., 2015). In the GoM, however, the submesoscale seasonality is further modulated by the riverine outflow. The freshwater fluxes enhance the submesoscale circulations by increasing lateral buoyancy gradients, yet suppress them by decreasing the mixed-layer depth (Molemaker et al., 2005; Thomas and Ferrari, 2008; McWilliams et al., 2009). The end result for the region offshore the shallow shelves is a slightly weaker submesoscale activity in winter than it would be in the absence of riverine forcing, minimal submesoscale activity recorded in spring, but stronger submesoscale dynamics and especially intense frontogenetic processes in summer and early fall, when the riverine outflow has reached offshore following the maximum discharge in late spring. Those submesoscale circulations further contribute to the lateral mixing of freshwater and to the offshore transport of the nitrogen of riverine origin, which fertilizes the otherwise nitrogen-limited offshore waters (Cardona et al., 2016), and are characterized by elevated vertical velocities (Zhong and Bracco, 2013; Zhong et al., 2017), with the potential to accelerate particle sinking. In 2012 the mixed-layer depth was shallower than the climatological average in March by about 4 m, and deeper than the climatological average in September by about 5 m. As a consequence, submesoscale circulations were even more abundant and intense in fall than in spring.

Figure 4 

Surface relative vorticity in April and September with superposed points of origin of particles. The color scale indicates relative vorticity normalized by the Coriolis parameter, ζ/f, at the ocean surface in April (left panels) and September (right panels), averaged over the time interval when the particles with sinking velocities equal to 30 m d–1 (a and b), 50 m d–1 (c and d) and 100 m d–1 (e and f) were also at the surface. Negative (positive) relative vorticity indicates anticyclonic (cyclonic) rotation. Points of origin of particles collected at GC600 (AT357) in black (green) in one ensemble member for each speed and collection period are also shown. DOI:

Submesoscale/mesoscale variability and catchment area of sinking particles

After describing the circulations that affect transport and mixing in the northern GoM in different seasons, we explore in detail the connectivity between upper layer circulation and flux at the depth of the sediment traps using Lagrangian tracers with vertical velocities of 20, 30, 50 and 100 m d–1, released from GC600 and AT357. Figure 4 identifies the location of the particles backtracked to the surface for one of the ensemble members performed and three sinking speeds (30, 50 and 100 m d–1; the distribution of particles for 20 m d–1 is qualitatively similar to that for 30 m d–1), superimposed on the corresponding vorticity field averaged over the interval when the particles reached the surface. Results indicate that the horizontal extent of sinking particles varied from tens to hundreds of kilometers. In April, a newly formed anticyclonic eddy propagated to the west, with the LC dominating the southeastern Gulf (Figure 2). The backtracked particles sinking at GC600 and AT357 show comparable horizontal dispersion independently of the sinking velocity when at the surface, with an east-west distribution above the seep for those ending at GC600 and a more north-south pattern for those reaching AT357 (e.g., Figure 4b). In October, however, particles reaching GC600 originated from a widely spread area if their sinking speed was 50 m d–1 or less; most of them can be traced to the periphery of the large anticyclonic LC eddy, with others spread along the continental slope to the northeast of the site. Particles as far as ~490 km apart reach the trap site at about the same time. Previous studies have reported large vertical velocities in or near the cyclonic cells of vorticity that surround anticyclonic eddies (Martin and Richards, 2001; McGillicuddy et al., 2007; Zhou et al., 2013). Such velocities are particularly strong and effective in presence of ageostrophic motions; i.e., when |ζ|/f is of order 1 (Zhong et al., 2017) as in the case of the September LC.

Tracers sinking at AT357, on the other hand, were confined to a much smaller region when at the surface, limited to a maximal distance of ~240 km with low but positive vorticity, and were not directly impacted by the LC eddy (Figure 4b, 4d and 4f). Videos of the sinking particles, uploaded as Videos S1–S4, provide a three-dimensional view of the pathways taken by the particles in one ensemble member for most of the sinking speeds considered.

The impacts of riverine inputs on fluxes at depths

As typical for regions influenced by riverine inflows, the spatial and temporal variability of POC in the GoM is strongly affected by the discharge from Mississippi-Atchafalaya River system. As riverine waters are usually less abundant and more confined along the shelf in winter and early spring, they reach their maximum in April or May and spread more effectively offshore in summer and early fall (Morey et al., 2003a), aided by the winds and by the mesoscale eddies and the LC. Chl-a patterns averaged over the periods 20–30 March 2012 and 1–15 September 2012 (Figure 5) exemplify in the observations as well as in the model the role of cross-shore transport of nutrient-rich riverine water in the presence of a strong mesoscale field. In March and early April, the ‘background’ primary productivity (blue/light blue range in Figure 5) at both sites is on average two to three times higher than in September (see Figure 2 in Giering et al., 2018) following the February winter bloom that characterizes the GoM due to frequent wind-mixing events that bring nutrients into the euphotic layer (Muller-Karger et al., 2015). Additionally, particles that sink to the traps at AT357 and GC600 occupy at the surface regions that are comparable in terms of size and (relatively high) Chl-a concentrations.

Figure 5 

Averaged Chl-a distribution in late March and September. Averaged Color scale indicates Chl-a concentration (logarithm) over the periods of 20–30 March 2012 (a) and 1–15 September 2012 (b), when the majority of particles collected in the traps were likely at the ocean surface according to the model results. Trap locations GC600 and AT357 are indicated by black and green plus symbols, respectively. Chl-a data were obtained from the NASA Ocean color portal. DOI:

In contrast, in the first two weeks of September, a southeastward low-salinity extension wraps around the LC eddy, as captured in both the Chl-a (Figure 5) and SSS maps (Figure 3). Abundant nutrients as well as organic matter carried by the intrusion of riverine water were likely to cover the area from which some of the particles that sank to GC600 may have originated. In the model runs, the remaining particles sinking into this trap came directly from the shelf, also a nutrient-rich area in the real ocean, or from other portions of the LC eddy circulation cell, which are characterized by strong vertical velocities. The coastal productive waters, the freshwater plume, or strongly ageostrophic vorticity, however, do not intersect the region where the particles that were collected at the model grid point of AT357 originated.

Figure 6 quantifies the riverine contributions for particles collected by the traps in April and October through the averaged probability density function (PDF) of the salinity at their surface positions in March and September for the seven ensemble members at sinking speeds of 30 m d–1 and 50 m d–1. The September–October histogram clearly displays a low-salinity tail (<34.5) only for the particles that end at GC600. In March–April PDFs are nearly identical for the two traps.

Figure 6 

Probability density functions of salinity at the particle surface locations. The average probability density functions (PDFs) of salinity at the points of origin of the particles reaching the traps are shown for April (a) and October (b). PDFs were calculated considering all particles (7 ensemble members per case) with sinking velocities of 30 m d–1 and 50 m d–1. DOI:

Lateral advection and sinking time

The cone of influence of sinking particles in the GoM amidst its mesoscale circulation was further evaluated through the PDFs of the lateral displacement of the point of origin of the particles at the surface with respect to the trap location (Figure S4). The cone of influence for fast-sinking particles (e.g., 100 m d–1) is observed to peak at about 70 km and increase to 200 km for particles sinking with speed around 20 m d–1. Some of the slowest particles, however, could be backtracked to regions as far as ~350 km. The impact of the LC eddies is especially evident at GC600 in September–October, where the distribution was centered around 300 km. Figure 7 conceptually shows these origins and pathways of simulated particles from the ocean surface to the depth of the traps in the presence of the LC and LC eddies in spring and fall 2012.

Figure 7 

Conceptual representation of transport of sinking particles in spring (a) and fall (b). Surface and bottom color shadings indicate sea surface height (SSH) over the whole domain and bathymetry along the northern shelf, respectively. The large black and green circles indicate the locations of sediment traps deployed at about 120 m above the bottom at GC600 and AT357. Small dots represent particles collected at GC600 (black) and AT357 (green). The red shadings indicate the anticyclonic (clockwise-rotating) Loop Current (LC) and LC eddies that strongly influence the upper circulation and transport (above ~1000 m) in the Gulf of Mexico. The black (green) shaded ellipses show the areas of origin of particles sinking at GC600 (AT357); yellow stripes indicate highly productive areas. Black arrow points to the outlet of the Mississippi River. According to the model results (Figure 2), in Spring 2012 (a) particles collected at traps originated from confined, common areas to the north of the traps far from large mesoscale features (an LC eddy was to the southwest of trap locations and the LC to the southeast of them). In Fall 2012 (b) a large LC eddy in the center of the domain dominated the circulation, and particles collected at GC600 originated from the eddy periphery and from inshore, river-influenced high-productive waters. Particles backtracked from AT357 had again a more local origin above the trap. DOI:

In terms of the time required for particles to reach the traps from the surface, the lower the sinking speed, the broader the distributions are, independently of the season (Figure 8). Slowly sinking particles can be stirred laterally more efficiently, especially in the upper 100–150 m of the water column, where mesoscale and submesoscale features are most intense, and can sample different circulation features. Noticeably in all cases particles have a large probability of reaching the bottom faster than would be expected by their sinking speed alone following a straight line. Submesoscale-induced horizontal convergence near the surface and in the mixed layer contributes to the aggregation of particles in regions of strong horizontal convergence (i.e., submesoscale circulations) that, on average, are characterized predominantly by downwelling velocities in the upper portion of the water column, accelerating their descent. Figure 9 quantifies this effect by displaying the distribution of the fluid vertical velocities sampled by the particles for all tracers with sinking speed of 30 m d–1 through their descent in the upper 200 m of the water column during the period of September–October. Downward (positive) velocities dominate the distributions independently of the site. Greater values are found for particles collected at GC600 that were concentrated along the convergent side of the circulation cell around the LC eddy (Zhong et al., 2012; 2017). The initial aggregation and sinking in one ensemble member for most cases is also clearly visualized in the videos uploaded as Videos S1–S4.

Figure 8 

Probability density functions of the particle residence time. The average probability density functions (PDFs) of the time required for particles to transit from the ocean surface to the depth of the traps are shown for the four different sinking speeds considered. Each PDF represents the average over all ensemble members. The time that would be required to reach the traps along a straight line only through vertical sinking, assuming no flow (ocean) vertical velocity, is indicated by vertical lines. DOI:

Figure 9 

Probability density functions of flow vertical velocities in the upper 200 m in October. The average probability density functions (PDFs) of the flow vertical velocities felt by particles while transiting in the upper 200 m of the water column are shown for each of the seven ensemble members separately in the October case with sinking speed of 30 m d–1 for a) AT357 and b) GC600. Positive values indicate downward vertical velocities (and vice versa). DOI:

Summary and conclusions

The horizontal and vertical transport of sinking particles in the northern GoM was investigated using regional ocean model simulations at 1-km resolution. Our numerical experiments were performed to help interpret sediment trap data collected at two adjacent, offshore locations in the northern Gulf. POC fluxes at the two trap sites varied appreciably temporally and spatially, suggesting that the sources of POC were highly variable. One trap, at AT357, was slightly shallower and less prone to be directly impacted by LC eddies due to its proximity to the Mississippi Slope. The POC concentrations in this trap appear to be influenced at the first order by the riverine discharge, with a significant correlation between the time series. The other trap, at GC600, is more often near the path of an LC eddy and does not co-vary with the riverine discharge.

The backtracking analysis of about 28,000 trajectories of particles sinking at four different velocities was performed to investigate the source of deep fluxes at the two trap sites at the end of April and beginning of October 2012. Our results suggest that in the GoM the variability of particle fluxes at depth is controlled by mesoscale eddies and the convergence induced by submesoscale circulations, as well as the seasonal cycling of primary productivity, riverine forcing and winds. In the northern Gulf primary productivity is generally highest in February, substantial through spring and lowest in summer and early fall. Riverine forcing is usually greatest in April–May and lowest in late fall, but highly variable on interannual scales. The winds, jointly with the mesoscale structures, determine the cross-shore transport of freshwater away from the shelf, which is commonly maximal in summer and early fall, and reduced in winter and early spring. In March–April, when the LC and a shed eddy dominated the GoM circulation, our model results suggest that particulates collected near the bottom at GC600 and AT357 originated from a cyclonic area of about 360 km between the two anticyclonic structures. In October the particles collected at GC600 originated from the shelf and from the cyclonic circulation cell around the LC eddy boundary that had entrained water from the Atchafalaya River. Both areas were characterized by low salinity, indicative of riverine water and high nutrient and particulate concentrations, in agreement with the high POC fluxes observed. By contrast, the particles backtracked from AT357 originated more locally, in an area that lacked strong vertical velocities and river influence, and that was far away from the LC eddy.

Our modeling experiment revealed that the combination of processes induced by mesoscale and submesoscale circulations and cross-shore transport of riverine inputs can result in POC fluxes in offshore waters that are highly variable in both space and time as quantified through probability density functions. Mesoscale circulations control size and shape of the catchment area, as well as the riverine water content and therefore productivity (nutrient influx). Submesoscale circulations, on the other hand, influence the overall time required for the marine snow to reach the ocean bottom, especially when particle sinking velocities are 50 m d–1 or less by converging particles within the mixed layer in regions where downward velocities are predominant and accelerating their descent. If the cone of influence for rapidly sinking particles (with sinking velocities of 100 m d–1 or higher) is generally limited to an area of a few tens of kilometers nearby the trap locations (Guidi and Stemmann, 2007), the cone of influence for particles with sinking speeds of 50 m d–1 or less can extend to several hundred kilometers depending on the details of the mesoscale circulation, in agreement with previous studies (Siegel et al., 1990; Siegel and Deuser, 1997; Waniek et al., 2000).

Another potential source of POC in the GoM, not considered in this work, is provided by the large number of natural hydrocarbon seeps across the entire northern GoM (MacDonald et al., 2002). Yearly averaged concentrations of POC are indeed higher at GC600, a well-known location of natural seeps, than at AT357, possibly due to physical and biological processes in response to the released hydrocarbons (e.g., bubble-induced upwelling flow and utilization of carbon and nutrients by plankton (D’souza et al., 2016).

Data Accessibility Statements

Data are publicly available through the Gulf of Mexico Research Initiative Information & Data Cooperative (GRIIDC) at (DOI:, DOI:, DOI:, DOI:, DOI: This is ECOGIG contribution 507.

Supplemental Files

The supplemental files for this article can be found as follows:


We thank two anonymous reviewers for their insightful comments that helped to improve the clarity of this paper.

Funding informations

The research was made possible by a grant from The Gulf of Mexico Research Initiative through the ECOGIG consortium.

Competing interests

The authors have no competing interests to declare.

Author contributions

  • Contributed to conception and design: GL, AB, UT
  • Contributed to numerical simulations: GL
  • Contributed to acquisition of field data: UT
  • Contributed to analysis and interpretation of data: GL, AB
  • Drafted and/or revised the article: GL, AB, UT
  • Approved the submitted version for publication: GL, AB, UT


  1. Alldredge, AL and Silver, MW. 1988. Characteristics, dynamics and significance of marine snow. Prog Oceanogr 20: 41–82. DOI: 10.1016/0079-6611(88)90053-5

  2. Alonso-González, IJ, Arístegui, J, Lee, C, Sanchez-Vidal, A, Calafat, A, et al. 2010. Role of slowly settling particles in the ocean carbon cycle. Geophys Res Lett 37(13). DOI: 10.1029/2010GL043827

  3. Alvera-Azcárate, A, Barth, A and Weisberg, RH. 2009. The surface circulation of the Caribbean Sea and the Gulf of Mexico as inferred from satellite altimetry. J Phys Oceanogr 39(3): 640–657. DOI: 10.1175/2008JPO3765.1

  4. Armstrong, RA, Lee, C, Hedges, JI, Honjo, S and Wakeham, SG. 2001. A new, mechanistic model for organic carbon fluxes in the ocean based on the quantitative association of POC with ballast minerals. Deep-Sea Res Pt II 49(1–3): 219–236 DOI: 10.1016/S0967-0645(01)00101-1

  5. Asper, VL. 1987. Measuring the flux and sinking speed of marine snow aggregates. Deep-Sea Res Pt I 34: 1–17. DOI: 10.1016/0198-0149(87)90117-8

  6. Barkan, R, McWilliams, JC, Molemaker, J, Choi, J, Srinivasan, K, et al. 2017b. Submesoscale dynamics in the northern Gulf of Mexico. Part II: Temperature-salinity relations and cross shelf transport processes. J Phys Oceanogr 47: 2347–2360. DOI: 10.1175/JPO-D-17-0040.1

  7. Barkan, R, McWilliams, JC, Shchepetkin, AF, Molemaker, J, Renault, L, et al. 2017a. Submesoscale dynamics in the northern Gulf of Mexico. Part I: Regional and seasonal characterization, and the role of river outflow. J Phys Oceanogr 47: 2325–2346. DOI: 10.1175/JPO-D-17-0035.1

  8. Biggs, DC and Müller-Karger, FE. 1994. Ship and satellite observations of chlorophyll stocks in interacting cyclone-anticyclone eddy pairs in the western Gulf of Mexico. J Geophys Res 99(C4): 7371. DOI: 10.1029/93JC02153

  9. Bracco, A, Choi, J, Joshi, K, Luo, H and McWilliams, JC. 2016. Submesoscale currents in the northern Gulf of Mexico: Deep phenomena and dispersion over the continental slope. Ocean Model 101: 43–58. DOI: 10.1016/j.ocemod.2016.03.002

  10. Bracco, A, Choi, J, Kurian, J and Chang, P. 2018. Vertical and horizontal resolution dependency in the model representation of tracer dispersion in the northern Gulf of Mexico. Ocean Model 122: 13–25. DOI: 10.1016/j.ocemod.2017.12.008

  11. Buesseler, KO, Antia, AN, Chen, M, Fowler, SW, Gardner, WD, et al. 2007. An assessment of the use of sediment traps for estimating upper ocean particle fluxes. J Mar Res 65: 345–416. DOI: 10.1357/002224007781567621

  12. Buesseler, KO, Lamborg, C, Cai, P, Escoube, R, Johnson, R, et al. 2008. Particle fluxes associated with mesoscale eddies in the Sargasso Sea. Deep-Sea Res Pt II 55: 1426–1444. DOI: 10.1016/j.dsr2.2008.02.007

  13. Callies, J, Ferrari, R, Klymak, JM and Gula, J. 2015. Seasonality in submesoscale turbulence. Nat Commun 6: 6862. DOI: 10.1038/ncomms7862

  14. Capet, X, Campos, EJ and Paiva, AM. 2008. Submesoscale activity over the Argentinian shelf. Geophys Res Lett 35(15). DOI: 10.1029/2008GL034736

  15. Cardona, Y, Ruiz-Ramos, DV, Baums, IB and Bracco, A. 2016. Potential connectivity of coldwater black coral communities in the northern Gulf of Mexico. PLOS One . 11(5): e0156257. DOI: 10.1371/journal.pone.0156257

  16. Chelton, DB, Gaube, P, Schlax, MG, Early, JJ and Samelson, RM. 2011. The influence of nonlinear mesoscale eddies on near-surface oceanic chlorophyll. Science 334: 328–332. DOI: 10.1126/science.1208897

  17. Dagg, MJ and Breed, GA. 2003. Biological effects of Mississippi River nitrogen on the northern Gulf of Mexico – a review and synthesis. J Mar Syst 43: 133–152. DOI: 10.1016/j.jmarsys.2003.09.002

  18. De La Rocha, CL and Passow, U. 2007. Factors influencing the sinking of POC and the efficiency of the biological carbon pump. Deep-Sea Res Pt II 54: 639–658. DOI: 10.1016/j.dsr2.2007.01.004

  19. Del Castillo, CE, Coble, PG, Conmy, RN, Muller-Karger, FE, Vanderbloemen, L, et al. 2001. Multispectral in situ measurements of organic matter and chlorophyll fluorescence in seawater: Documenting the intrusion of the Mississippi River plume in the West Florida Shelf. Limnol Oceanogr 46(7): 1836–1843. DOI: 10.4319/lo.2001.46.7.1836

  20. D’souza, NA, Subramaniam, A, Juhi, AR, Hafez, M, Chekalyuk, A, et al. 2016. Elevated surface chlorophyll associated with natural oil seeps in the Gulf of Mexico. Nat Geosci 9: 215–218. DOI: 10.1038/ngeo2631

  21. Dunn, DD. 1996. Trends in nutrient inflows to the Gulf of Mexico from streams draining the conterminous United States. 1972–1993. Austin (TX): US Geological Survey. Water-Resources Investigations Report no. 96-4113.

  22. Elliott, BA. 1982. Anticyclonic Rings in the Gulf of Mexico. J Phys Oceanogr 12: 1292–1309. DOI: 10.1175/1520-0485(1982)012<1292:ARITGO>2.0.CO;2

  23. Giering, SLC, Yan, B, Sweet, J, Asper, V, Diercks, A, et al. 2018. The ecosystem baseline for particle flux in the Northern Gulf of Mexico. Elem Sci Anth 6(1): 6. DOI: 10.1525/elementa.264

  24. Gilbes, F, Muller-Karger, FE and Del Castillo, CE. 2002. New evidence for the West Florida Shelf Plume. Cont Shelf Res 22: 2479–2496. DOI: 10.1016/S0278-4343(02)00102-4

  25. Guidi, L and Stemmann, L. 2007. Vertical distribution of aggregates (>110 um) and mesoscale activity in the northeastern Atlantic: Effects on the deep vertical export of surface carbon. Limnol Oceanogr 52(2): 7–18. DOI: 10.4319/lo.2007.52.1.0007

  26. Kiko, R, Biastoch, A, Brandt, P, Cravatte, S, Hauss, H, et al. 2017. Biological and physical influences on marine snow fall at the equator. Nat Geosci 10: 852–859. DOI: 10.1038/ngeo3042

  27. Lehahn, Y, d’Ovidio, F, Lévy, M, Amitai, Y and Heifetz, E. 2011. Long range transport of a quasi isolated chlorophyll patch by an Agulhas ring. Geophys Res Lett 38(16). DOI: 10.1029/2011GL048588

  28. Lévy, M, Ferrari, R, Franks, PJS, Martin, AP and Rivière, P. 2012a. Bringing physics to life at the submesoscale. Geophys Res Lett 39(14). DOI: 10.1029/2012GL052756

  29. Lévy, M, Iovino, D, Resplandy, L, Klein, P, Madec, G, et al. 2012b. Large-scale impacts of submesoscale dynamics on phytoplankton: Local and remote effects. Ocean Model 43–44: 77–93. DOI: 10.1016/j.ocemod.2011.12.003

  30. Lévy, M, Klein, P and Ben Jelloul, M. 2009. New production stimulated by high-frequency winds in a turbulent mesoscale eddy field. Geophys Res Lett 36(16). DOI: 10.1029/2009GL039490

  31. Lohrenz, SE, Fahnenstiel, GL, Redalje, DG, Lang, GA, Chen, X, et al. 1997. Variations in primary production of northern Gulf of Mexico continental shelf waters linked to nutrient inputs from the Mississippi River. Mar Ecol Prog Ser 155: 435–454. DOI: 10.3354/meps155045

  32. Luo, H, Bracco, A, Cardona, Y and McWilliams, JC. 2016. Submesoscale circulation in the northern Gulf of Mexico: Surface processes and the impact of the freshwater river input. Ocean Model 101: 68–82. DOI: 10.1016/j.ocemod.2016.03.003

  33. MacDonald, IR, Leifer, I, Sassen, R, Stine, P, Mitcheell, R and Guinasso, N. 2002. Transfer of hydrocarbons from natural seeps to the water column and atmosphere. Geofluids . 2: 95–107. DOI: 10.1046/j.1468-8123.2002.00023.x

  34. Mahadevan, A. 2016. The impact of submesoscale physics on primary productivity of plankton. Annu Rev Mar Sci 8: 161–184. DOI: 10.1146/annurev-marine-010814-015912

  35. Mahadevan, A, D’Asaro, E, Lee, C and Perry, MJ. 2012. Eddy-driven stratification initiates North Atlantic spring phytoplankton blooms. Science 337: 54–58. DOI: 10.1126/science.1218740

  36. Marchesiello, P, McWilliams, JC and Shchepetkin, A. 2003. Equilibrium structure and dynamics of the California Current System. J Phys Oceanogr 33: 753–783.

  37. Martin, AP. 2003. Phytoplankton patchiness: the role of lateral stirring and mixing. Prog Oceanogr 57: 125–174. DOI: 10.1016/S0079-6611(03)00085-5

  38. Martin, AP and Richards, KJ. 2001. Mechanisms for vertical nutrient transport within a North Atlantic mesoscale eddy. Deep-Sea Res Pt II 48: 757–773. DOI: 10.1016/S0967-0645(00)00096-5

  39. McGillicuddy, D, Anderson, LA, Bates, NR, Bibby, T, Buesseler, KO, et al. 2007. Eddy/wind interactions stimulate extrordinary mid-ocean plankton blooms. Science 316: 1021–1026. DOI: 10.1126/science.1136256

  40. McWilliams, JC. 2008. Fluid dynamics at the margin of rotational control. Environ Fluid Mech 8(5–6): 441–449. DOI: 10.1007/s10652-008-9081-8

  41. McWilliams, JC, Molemaker, MJ and Olafsdottir, EI. 2009. Linear fluctuation growth during frontogenesis. J Phys Oceanogr 39(12): 3111–3129. DOI: 10.1175/2009JPO4186.1

  42. Mellor, GL and Yamada, T. 1982. Development of a turbulence closure model for geophysical fluid problems. Rev Geophys 20: 851–875. DOI: 10.1029/RG020i004p00851

  43. Mensa, JA, Garraffo, Z, Griffa, A, Özgökmen, TM, Haza, A, et al. 2013. Seasonality of the submesoscale dynamics in the Gulf Stream region. Ocean Dyn 63(8): 923–941. DOI: 10.1007/s10236-013-0633-1

  44. Molemaker, MJ, McWilliams, JC and Yavneh, I. 2005. Baroclinic instability and loss of balance. J Phys Oceanogr 35: 1505–1517. DOI: 10.1175/JPO2770.1

  45. Morey, SL, Martin, PJ, O’Brien, JJ, Wallcraft, AA and Zavala-Hidalgo, J. 2003a. Export pathways for river discharged fresh water in the northern Gulf of Mexico. J Geophys Res 108(C10). DOI: 10.1029/2002JC001674

  46. Morey, SL, Schroeder, WW, O’Brien, JJ and Zavala-Hidalgo, J. 2003b. The annual cycle of riverine influence in the eastern Gulf of Mexico basin. Geophys Res Lett 30(16). DOI: 10.1029/2003GL017348

  47. Muller-Karger, FE, Smith, JP, Werner, S, Chen, R, Roffer, M, et al. 2015. Natural variability of surface oceanographic conditions in the offshore Gulf of Mexico. Prog Oceanogr 134: 54–76. DOI: 10.1016/j.pocean.2014.12.007

  48. Neuer, S, Ratmeyer, V, Davenport, G, Fischer, G and Wefer, G. 1997. Deep water particle flux in the Canary Islands region: seasonal trends in relation to long term satellite derived pigment data and lateral sources. Deep-Sea Res Pt I 44: 1451–1466. DOI: 10.1016/S0967-0637(97)00034-4

  49. Omand, MM, D’Asaro, EA, Lee, CM, Perry, MJ, Briggs, N, et al. 2015. Eddy-driven subduction exports particulate organic carbon from the spring bloom. Science 348(6231): 222–225. DOI: 10.1126/science.1260062

  50. Passow, U and Carlson, CA. 2012. The biological pump in a high CO2 world. Mar Ecol Prog Ser 470: 249–271. DOI: 10.3354/meps09985

  51. Passow, U, Ziervogel, K, Asper, V and Diercks, A. 2012. Marine snow formation in the aftermath of the Deepwater Horizon oil spill in the Gulf of Mexico. Environ Res Lett 7(3). DOI: 10.1088/1748-9326/7/3/035301

  52. Poli, P, Healy, SB and Dee, DP. 2010. Assimilation of Global Positioning System radio occultation data in the ECMWF ERA-interim reanalysis. Q J R Meteorol Soc 13: 1970–1990. DOI: 10.1002/qj.722

  53. Qiu, B, Chen, S, Klein, P, Sasaki, H and Sasai, Y. 2014. Seasonal mesoscale and submesoscale eddy variability along the North Pacific Subtropical Countercurrent. J Phys Oceanogr 44(12): 3079–3098. DOI: 10.1175/JPO-D-14-0071.1

  54. Renault, L, Molemaker, MJ, McWilliams, JC, Shchepetkin, AF, Lemarie, F, et al. 2016. Modulation of wind work by oceanic current interaction with the atmosphere. J Phys Oceanogr 46: 1685–1704. DOI: 10.1175/JPO-D-15-0232.1

  55. Roberts, HH, Feng, D and Joye, SB. 2010. Cold-seep carbonates of the middle and lower continental slope, northern Gulf of Mexico. Deep-Sea Res Pt II 57(21): 2040–2054. DOI: 10.1016/j.dsr2.2010.09.003

  56. Sandwell, DT and Smith, WHF. 1997. Marine gravity anomaly from Geosat and ERS-1 satellite altimetry. J Geophys Res 102: 10,039–10,054

  57. Schlag, ZR and North, EW. 2012. Lagrangian TRANSport model (LTRANS v.2) User’s Guide. University of Maryland Center for Environmental Science, Horn Point Laboratory. Cambridge, MD. 183. (Available at:

  58. Siegel, DA and Deuser, WG. 1997. Trajectories of sinking particles in the Sargasso Sea: modeling of statistical funnels above deep-ocean sediment traps. Deep-Sea Res Pt I 44: 1519–1541. DOI: 10.1016/S0967-0637(97)00028-9

  59. Siegel, DA, Granata, TC, Michaels, AF and Dickey, TD. 1990. Mesoscale eddy diffusion, particle sinking, and the interpretation of sediment trap data. J Geophys Res 95(C4): 5305. DOI: 10.1029/JC095iC04p05305

  60. Sikiric, M, Janekovic, I and Kuzmic, M. 2009. A new approach to bathymetry smoothing in sigma-coordinate ocean models. Ocean Model 29(2): 128–136. DOI: 10.1016/j.ocemod.2009.03.009

  61. Simon, M, Grossart, H, Schweitzer, B and Ploug, H. 2002. Microbial ecology of organic aggregates in aquatic ecosystems. Aquat Microb Ecol 28: 175–211. DOI: 10.3354/ame028175

  62. Son, YB, Gardner, WD, Mishonov, AV and Richardson, MJ. 2009. Multispectral remote-sensing algorithms for particulate organic carbon (POC): The Gulf of Mexico. Remote Sens Environ 113(1): 50–61. DOI: 10.1016/j.rse.2008.08.011

  63. Sturges, W and Leben, R. 2000. Frequency of ring separations from the Loop Current in the Gulf of Mexico. J Phys Oceanogr 30: 1814–1819. DOI: 10.1175/1520-0485(2000)030<1814:FORSFT>2.0.CO;2

  64. Thomas, L and Ferrari, R. 2008. Friction, frontogenesis, and the stratification of the surface mixed layer. J Phys Oceanogr 38(11): 2501–2518. DOI: 10.1175/2008JPO3797.1

  65. Toner, M, Kirwan, AD, Poje, AC, Kantha, LH, Muller-Karger, FE, et al. 2003. Chlorophyll dispersal by eddy-eddy interactions in the Gulf of Mexico. J Geophys Res 108(C4). DOI: 10.1029/2002JC001499

  66. Vukovich, FM. 1988. Loop Current boundary variations. J Geophys Res 93(C12): 15585. DOI: 10.1029/JC093iC12p15585

  67. Vukovich, FM. 2007. Climatology of ocean features in the Gulf of Mexico using satellite remote sensing data. J Phys Oceanogr 37(3): 689–707. DOI: 10.1175/JPO2989.1

  68. Waniek, J, Koeve, W and Prien, RD. 2000. Trajectories of sinking particles and the catchment areas above sediment traps in the northeast Atlantic. J Mar Res 58: 983–1006. DOI: 10.1357/002224000763485773

  69. Weisberg, RH, Black, BD and Li, Z. 2000. An upwelling case study on Florida’s west coast. J Geophys Res: Oceans 105(C5): 11459–11469. DOI: 10.1029/2000JC900006

  70. Zhong, Y and Bracco, A. 2013. Submesoscale impacts on horizontal and vertical transport in the Gulf of Mexico. J Geophys Res: Oceans 118(10): 5651–5668. DOI: 10.1002/jgrc.20402

  71. Zhong, Y, Bracco, A, Tian, J, Dong, J, Zhao, W, et al. 2017. Observed and simulated vertical pump of an anticyclonic eddy in the South China Sea. Sci Rep , 44011. DOI: 10.1038/srep44011

  72. Zhong, Y, Bracco, A and Villareal, TA. 2012. Pattern formation at the ocean surface: Sargassum distribution and the role of the eddy field. Limnol Oceanogr 2(1): 12–27. DOI: 10.1215/21573689-1573372

  73. Zhou, K, Dai, M, Kao, SJ, Wang, L, Xiu, P, et al. 2013. Apparent enhancement of 234Th-based particle export associated with anticyclonic eddies. Earth Planet Sci Lett 381: 198–209. DOI: 10.1016/j.epsl.2013.07.039

comments powered by Disqus