Content uploaded by Graeme C Hays
Author content
All content in this area was uploaded by Graeme C Hays
Content may be subject to copyright.
MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 353: 289– 301, 2008
doi: 10.3354/meps07118 Published January 17
INTRODUCTION
For animals travelling in a moving medium, the
movement of the medium may have profound impacts
on the animal’s trajectory. Classic examples concern
the impact of wind on the journeys of birds and of
ocean currents on the movement of marine animals,
especially as regards larval dispersion (Dingle 1996,
Alerstam et al. 2003). However, currents may also
affect nektonic animals capable of strong active swim-
ming, like sea turtles (Luschi et al. 2003a). Turtle
hatchlings are advected around entire ocean basins by
currents (Musick & Limpus 1997), but adult turtles
could also be affected by current circulation in their
movements (e.g. Girard et al. 2006). This applies even
to the largest species of turtle, the pelagic-dwelling
leatherback turtle Dermochelys coriacea, whose pref-
erences for patchily distributed pelagic prey cause it
to wander over large regions while feeding on
macroplankton. Indeed, satellite tracking studies have
shown that leatherbacks carry out extensive oceanic
movements, following complex routes showing a high
degree of inter- and intra-individual variation (Mor-
reale et al. 1996, Eckert & Sarti 1997, Luschi et al.
© Inter-Research 2008 · www.int-res.com*Corresponding author. Email: pluschi@biologia.unipi.it
Influence of ocean currents on long-distance
movement of leatherback sea turtles in the
Southwest Indian Ocean
Paolo Lambardi1, Johann R. E. Lutjeharms2, Resi Mencacci1, Graeme C. Hays3,
Paolo Luschi1,*
1Dipartimento di Biologia, Università di Pisa, Via A. Volta 6, 56126 Pisa, Italy
2Department of Oceanography, University of Cape Town, Rondebosch 7700, South Africa
3Department of Biological Sciences, Institute of Environmental Sustainability, University of Wales Swansea, Singleton Park,
Swansea SA2 8PP, UK
ABSTRACT: Leatherback turtles Dermochelys coriacea spend most of their life in oceanic environ-
ments, whose physical and biological characteristics are primarily forged by sea current circulation.
Water mass movements can mechanically act on swimming turtles, thus determining their routes, and
can differentially distribute their planktonic prey. By integrating satellite tracking data with contem-
poraneous remote-sensing information, we analysed the post-nesting journeys of 9 leatherbacks with
respect to oceanographic surface conditions. Tracked turtles showed large variations in migration
routes and in final destinations, apparently without heading for specific foraging areas. Their com-
plex tracks spread over wide regions around South Africa. Leatherbacks were greatly influenced by
the currents encountered during their movements, with their trajectories displaying curves or revolu-
tions in the presence of (and in accordance with) rotating water masses. An impressive similarity was
observed between large parts of the turtle routes and those of surface drifters tracked in the same
regions. Finally, leatherbacks remained associated for long periods with specific oceanographic
features, which most probably offered them profitable foraging opportunities. These results agree
with previous findings in showing a strong influence of oceanic currents and mesoscale features on
the movements of South African leatherbacks, and additionally identify the role of current-related
features in causing the observed route variability and in determining high-quality foraging hotspots
for leatherbacks moving in the ocean.
KEY WORDS: Satellite telemetry · Dermochelys · Remote sensing · Oceanography · Current drift
Resale or republication not permitted without written consent of the publisher
Mar Ecol Prog Ser 353: 289–301, 2008
2003b, Ferraroli et al. 2004, Hays et al. 2004a,b, 2006,
James et al. 2005, Eckert 2006).
The oceanic areas visited by foraging turtles are
mainly characterised by sea currents and related fea-
tures, which can influence leatherback feeding-related
movements in 2 main ways: (1) currents exert mechan-
ical actions on swimming animals, thus affecting the
shape of the route taken; (2) oceanic circulation is
responsible for differential distributions of food
resources in the ocean, determining the distribution of
planktonic primary producers, and thus also of animals
at higher trophic levels, including top predators like
leatherbacks. In particular, eddies may concentrate
nutrients and organisms, and thus represent patches of
high prey abundance targeted by foraging turtles. By
integrating satellite tracking findings with contempo-
raneous oceanographic data, the influence of these
processes on ocean-moving leatherbacks can be reli-
ably investigated. For instance, analysing the relation-
ship between the turtle movement direction and the
water flow in that area can highlight the marked influ-
ence of currents on turtle movements (e.g. Gaspar et
al. 2006). Similarly, if specific oceanographic elements
provide important feeding opportunities for turtles,
prolonged sojourns in those areas would be expected.
For the present study, we compiled tracking data
from 9 leatherback turtles tracked over 8 yr during
their wide-range oceanic movements in the SW Indian
Ocean, and we integrated these with available oceano-
graphic information. Three of these routes have been
discussed in a previous paper (Luschi et al. 2003b), in
which only sea current effects on the general route
shapes were considered. Their inclusion in the present
study allowed us to analyse movements performed
over several years and varying in course and destina-
tions, and thus to also investigate whether, and to
which extent, intra- and inter-individual variations in
migration patterns are linked to the spatio-temporal
variability in oceanographic conditions of the region.
MATERIALS AND METHODS
Turtles and transmitters. Nine female leatherbacks
nesting in the Maputaland Marine Reserve, on the
eastern coast of South Africa, were followed through
the Argos system during their post-nesting movements
between 1996 and 2003. Three different models of
transmitters, produced by Telonics and by the Sea
Mammal Research Unit (University of St. Andrews,
UK), were used. They were programmed with different
duty cycles (Table 1) and placed on the carapace with
harnesses (Luschi et al. 2003b). The Argos system pro-
vided location data classified into 6 accuracy levels,
and the routes were reconstructed using all fixes and
filtering out locations on land or producing ground
speed values >10 km h–1 (a threshold estimated from
high-accuracy locations only).
Study area: physical setting. The SW Indian Ocean
is an optimal region in which to investigate the behav-
iour of marine animals in relation to ocean currents.
While the water flow of this region is dominated by the
greater Agulhas Current system (Fig. 1a), a wide range
of current types, water speeds, scales of motion and
potential food sources are found here (Lutjeharms
2006).
The most prominent current in the SW Indian
Ocean is the Agulhas Current, which flows along the
east coast of South Africa following the continental
shelf edge quite closely with a very stable trajectory.
The current starts somewhere between the cities of
Durban and Maputo, but the location of its northern-
290
Turtle Tracking period Transmitter Duty cycle Percentage of
model route determined
by currents
Lara 16 Jan–18 May 1996 ST–14 Continuously on 84
Nemi 31 Jan –30 Sept 1999 ST–6 Continuously on 78
Sandra 31 Jan–11 Sept 1999 ST–6 Continuously on 69
Resi 30 Jan–16 Jun 2000 ST– 6 5 d off/1 d on 84
Elena 2 Feb–3 Mar 2000 ST–6 Continuously on NA
4 Mar–2 Apr 2000 5 d off/1 d on
Ronel 13 Feb–15 Mar 2001 ST– 6 Continuously on 40
16 Mar–20 May 2001 5 d off/1 d on
Alice 13 Feb –16 Mar 2001 ST–6 Continuously on 72
17 Mar– 8 Jun 2001 5 d off/1 d on
Imola 13 Jan–30 Jan 2002 SRDL Continuously on 93
Sara 29 Jan–16 Jul 2003 SRDL Continuously on 65
Table 1. Summary information for the 9 turtles tracked between 1996 and 2003. NA: estimation not available since ‘Elena’
remained in coastal waters where no reliable sea surface height anomaly (SSHA) information was accessible
Lambardi et al.: Influence of ocean current on sea turtle movement
most expression is uncertain and quite variable (Flem-
ming & Hay 1988). The current is 60 to 100 km wide
at the sea surface (Beal & Bryden 1999), and speeds in
its core can exceed 7.2 km h–1, decreasing with dis-
tance offshore. Its waters are distinctly warmer and
saltier than the ambient water masses, and are gener-
ally oligotrophic with strips of higher productivity at
its borders (Lutjeharms 2006).
Southwest of the continent, the current
retroflects with most of its water return-
ing eastward to the Indian Ocean along
the Subtropical Convergence as the
Agulhas Return Current (Fig. 1a) (Lutje-
harms & Ansorge 2001). At the western
termination of the Agulhas Current,
huge rings of water are shed by the
occlusion of the retroflection loop, subse-
quently drifting into the South Atlantic
Ocean at 5 to 15 km d–1. The Agulhas
Return Current exhibits large meanders,
and its surface speed decreases with dis-
tance eastward. Events of high produc-
tivity are spasmodically found at the
Subtropical Convergence (Llido et al.
2005). At 60 to 70° E, the effect of the
Agulhas Return Current is lost and the
flow along the Subtropical Convergence
is much weaker.
The Benguela Current (Fig. 1a) on the
western side of southern Africa has
entirely different characteristics. It rep-
resents the weak eastern boundary cur-
rent of the SE Atlantic Ocean: estimates
of the mean surface drift give average
drift rates of 5 to 15 km d–1 (Wedepohl et
al. 2000). Adjacent to the Benguela Cur-
rent proper lies a wind-driven coastal
upwelling system (Shannon & Nelson
1996), a highly productive region
extending up to the shelf edge, with
occasional filaments of cold, nutrient-
rich water spreading far offshore (Lutje-
harms et al. 1991).
Finally, the region offshore SE Africa is
characterised by the passage of large
eddies formed in the narrows of the
Mozambique Channel and drifting pole-
ward (de Ruijter et al. 2002). These deep-
sea eddies form an important component
of the circulation in the area and may be
joined by similarly intense eddies com-
ing from east of Madagascar (e.g. de
Ruijter et al. 2003).
This assemblage of an intense western
boundary current, a weak eastern
boundary current adjacent to a coastal upwelling, and
a host of eddies of different dimensions, intensities and
senses of rotation represents nearly the full range of
ocean features that marine animals could encounter
anywhere in the world.
Remote sensing and oceanographic data. The
reconstructed turtle routes were analysed in relation to
contemporaneous satellite remote-sensing data on sea
291
Fig. 1. (a) Schematic map of the general circulation of surface currents in the
SE Atlantic and SW Indian oceans. Redrawn from Richardson et al. 2003.
(b) Routes of 21 surface drifters tracked as they passed within the region
delimited by the black box. Different colours highlight distinct movement pat-
terns. Pink: drifters remaining at low latitudes; green: drifters circuiting in
areas off the Agulhas mainstream; black: drifters tracked only within the
Agulhas mainstream; blue: drifters entering the Atlantic Ocean; red: drifters
being captured by the Agulhas Return Current (thick track: drifter shown in
Fig. 4). The duration of plotted trajectories ranges from 51 to 657 d (mean ± SE,
181.9 ± 35.1 d).White and green dots: positions of turtle nesting sites
Mar Ecol Prog Ser 353: 289–301, 2008
surface temperature (SST), sea surface height anom-
alies (SSHA) and chlorophyll (chl) aconcentration.
SST was derived from multi-channel data sensed
through the Advanced Very High Resolution Radiome-
ter (AVHRR) on board the NOAA 14 satellite (nomi-
nally 1 km spatial resolution). Data were gridded by
optimum interpolation by the Modular Ocean Data
Assimilation System and made available by the Naval
Research Laboratory (Stennis Space Center, Missis-
sippi, USA; http://www7320.nrlssc.navy.mil/altimetry).
SSHA images, derived from measurements made by
Jason, TOPEX/Poseidon, Geosat Follow-on, ERS-2 and
Envisat altimetry satellites, are available from the Col-
orado Center for Astrodynamic Research (Univer-
sity of Colorado, Boulder, Colorado, USA; http://argo.
colorado.edu/~realtime/gsfc_global-real-time_ssh/) in
the form of 10 d averages ending at the given date.
They reveal variations in course and speed of major
currents and identify the occurrence of mesoscale
oceanographic features, like eddies or filaments, while
SST images give indications of the general course of
the Agulhas Current. The integration of these 2 infor-
mation sources is particularly valuable for portraying
rapidly changing oceanographic features, such as
passing eddies.
Finally, ocean colour images were obtained from the
Sea-viewing Wide Field-of-view Sensor (SeaWiFS)
on board the Orb-View 2 satellite (SeaWiFS Project,
http://oceancolor.gsfc.nasa.gov/). They were used to
get indications of the chl aconcentration in the upper
layers of the water column and thus, their productivity.
Additionally, we considered the existing
observations of surface Lagrangian buoys
tracked with Argos in the same region as
turtles. Drifter data were obtained from
the Atlantic Oceanographic Meteoro-
logical Laboratory (www.aoml.noaa.gov/
phod/trinanes/xbt.html). These driftersare
equipped with a drogue that acts like an
underwater sail, helping them to move
within the current flow and minimising
wind drag (Stewart 2003). In this way, they
provide reliable information on the circu-
lation of the upper layers of the water col-
umn (up to 100 m), where leatherbacks
spend most of their time (Sale et al. 2006).
To compare these routes to those of the
turtles, we chose all drifters (n = 21) that
passed through the region from 25 to
30° S, 32 to 36° E in 1996 to 2003 (see
Fig. 1b). We considered the tracks of these
drifters starting when they entered this
region and onwards. To evaluate their
speed while in the Agulhas mainstream,
the track segments running along the
eastern coast of South Africa were selected (n = 14).
The possibility of placing drifters concurrently with the
departure of turtles from nesting beaches would have
been ideal. However, in a current as geographically
stable as the Agulhas Current (Gründlingh 1983), non-
simultaneous drifters are reliable estimators for
Lagrangian motion. This is not completely true outside
of the current itself, where a wide range of circulation
characteristics are found.
GIS software (ArcView GIS 3.2) was used to plot the
routes of turtles and drifters and to superimpose the
turtle routes on the remote sensing images. To calcu-
late the percentage of each turtle route determined by
ocean currents (Table 1), we summed the length of the
various segments of each turtle route that were conser-
vatively estimated to derive from the direct influence
of current-related features (see next paragraph),
divided by the total length of the route.
RESULTS
General movement patterns
The tracks of the 9 leatherback turtles showed a
large degree of variation (Fig. 2), with a number of dif-
ferent movement patterns. A first distinction can be
made between the 4 turtles that remained at low lati-
tudes at least for several weeks and the 5 animals mov-
ing SW soon after leaving the nesting areas. Of these,
2 kept a course closely parallel to the South African
292
Fig. 2. Reconstructed routes of 9 leatherback turtles moving around
the southernmost part of the African continent. See text and Fig. 1 for
further details
Lambardi et al.: Influence of ocean current on sea turtle movement
coastline, while 3 headed more offshore,
in 2 cases spending long periods in
seemingly restricted areas. Of the 4 for-
mer turtles, 1 (‘Elena’) remained in
inshore waters for the whole tracking
period (2 mo), while the other 3 moved
offshore, where ‘Resi’ and ‘Alice’ per-
formed large loops. Further, other pat-
terns may also be discerned for the 2 tur-
tles entering the SE Atlantic Ocean and
for the 1 turning eastward.
Drifters passing off NE South Africa
displayed movements broadly similar to
those of the turtles (Fig. 1b). Three of the
21 selected buoys remained at latitudes
<30° S (pink routes), whereas the rest
generally moved SW. Of these, 6 dis-
played prolonged circuitous tracks over
a wide region off the Agulhas Current
mainstream (green routes), with 3 of
them reaching waters east of 35° E,
which were never frequented by tracked
turtles. Finally, of the 12 drifters that fol-
lowed the Agulhas Current mainstream,
4 entered the Atlantic basin and 3 were
captured by the Agulhas Return Current
(blue and red routes).
To investigate how the described gen-
eral patterns of movement of leatherback
turtles agree with the circulation pat-
terns of the region, it is convenient to
treat the different cases separately.
‘Elena’ was not considered for this analy-
sis, as no reliable SSHA information was
available for coastal regions.
Movements in the Agulhas Current proper
A rapid SW motion along the coast (Fig. 3) was the
type of movement most commonly displayed by our tur-
tles over the 8 yr period. It was usually undertaken soon
after the turtles left the nesting area, although ‘Resi’ and
‘Ronel’ did so only after several weeks of residence at
low latitudes. All turtles clearly moved in the same direc-
tion of the Agulhas Current main flow (Gründlingh
1983), depicted in Fig. 3 as a narrow ribbon of warm wa-
ter with temperatures over 24°C. The mean ± SE speed
for the 7 turtles in this leg was 79.0 ± 13.8 km d–1, far
in excess of those recorded in leatherbacks tracked in
other regions to date, either during inter-nesting (28.8 to
53.6 km d–1; Keinath & Musick 1993, Eckert 2002) or
post-nesting (34.0 to 71.2 km d–1; Duron-Dufrenne 1987,
Lutcavage et al. 2002, Ferraroli et al. 2004). While the
current contribution to these speeds is not known in any
case, this comparison clearly indicates that South African
turtles were substantially helped in their movements by
the rapid flow of the Agulhas Current.
Surface drifters moved along the same route as the
7 turtles at a comparable speed (76.3 ± 7.8 km d–1;
n = 14). Inter-annual variations were recorded in the
turtle speeds during this leg, which paralleled similar
variations in drifters (Fig. 3 insert). Apart from the gen-
eral downstream tendency, the turtle motion in the
Agulhas Current also included other indicative pat-
terns. At about the half-way point of their SW leg,
3 turtles made a seaward excursion at about 32° S
(‘Sara’, ‘Nemi’ and ‘Sandra’; Fig. 2), approximately at
the location of intermittent loss of water from the Agul-
has Current toward the central SW Indian Ocean
(Lutjeharms & van Ballegooyen 1988, Lutjeharms et al.
1992). Indeed, SST images (Fig. 3) clearly show that
such a loss occurred in February 1999 (when ‘Sandra’
and ‘Nemi’ made the excursion), while no direct evi-
dence is available for ‘Sara’ in 2003.
293
Fig. 3. South-westward oriented segments of the routes of 7 turtles superim-
posed on a sea surface temperature (SST) image recorded on 20 April 1999.
The position of the mainstream of the Agulhas Current is depicted as a ribbon
of warm water along the east coast of South Africa, which retroflects eastward
south of the continent. Insert: mean (±SE) speed of turtles and drifters tracked
along the Agulhas mainstream in different years. Speed was calculated as
the ratio between the distance covered within the Agulhas mainstream
and the time taken to cover this leg. Numbers indicate sample sizes. Other
explanations as in Fig. 1
Mar Ecol Prog Ser 353: 289–301, 2008
The similarity between the movement of drifters and
turtles may extend into the Agulhas Retroflection and
Return Current region. For instance, the tracks of
‘Lara’ and 1 drifter shown in Fig. 4 were both recorded
in 1996, but about 4 mo apart, and provide important
indications of how closely turtles are tied to the water
movement in the region south of the continent. A major
consistency was the tendency of the turtle and the
drifter to follow the water through the turbulent Agul-
has Retroflection and to emerge in the Agulhas Return
Current. When approaching the Retroflection region,
both the turtle and the drifter underwent extensive
anti-clockwise motion at about 22° E: ‘Lara’ once, and
the drifter twice. This can therefore be assumed to
have been the general location of the Agulhas
Retroflection at the time. Both then escaped from the
Retroflection carrying out gyrations in Agulhas rings to
the west of the Retroflection. Subsequently, they were
both carried away in the Agulhas Return Current,
albeit during rather different periods (April vs. October
1996).
On the whole, when tracked turtles found them-
selves within the Agulhas Current mainstream, their
movement fit closely with the SW current flow. On the
other hand, the Agulhas mainstream possibly offered
few feeding opportunities to leatherbacks, as no clear
signal of enhanced productivity in this region was
identifiable from ocean colour images. During their
stay in this area, the turtles indeed frequented waters
where the chl adensity was always below 0.2 mg m–3.
The inter-oceanic shift
Two turtles (‘Sandra’ and ‘Sara’) made an inter-oceanic
shift to the Atlantic Ocean (Fig. 2). In this case, surface
currents also played a major role in determining this
movement, highlighted as well by the frequent occur-
rence of such shifts in surface drifters (Fig. 1b). Both tur-
tles, which were tracked in different years, were clearly
in the core of the Agulhas Current during the last part of
their journey in the Indian Ocean (see Fig. 5a for ‘Sara’),
moving at rates of 62.3 and 119.6 km d–1, respectively,
between the longitudes of Port Elizabeth and Cape
Town. As evident in the SST images, they then moved
out of the main current from the western edge of the Ag-
ulhas Retroflection, without being retained in the
Retroflection loop. During this shift, their movements
were in accordance with the main surface circulation in
the area. SSHA images show that the region directly
west of the Retroflection was indeed filled with a collec-
tion of eddies (Fig. 5b for ‘Sara’), which were moving NW
at the time. The influence of such features on turtle
movements is shown comprehensively by the track of
‘Sara’ (Fig. 5b), which followed the contours of SSHA
fairly well. On escaping from the Agulhas Retroflection,
she was caught by an anti-cyclonic eddy that made
her track loop anti-clockwise (around 13 March). A sec-
ond anti-cyclonic eddy subsequently led her to culmi-
nate in the farthest eastward divergence from a straight
meridional track (16 March), before leading her to the
westernmost excursion on 21 March.
Entering the Atlantic Ocean allowed
‘Sandra’ and ‘Sara’ to access the upwelling
regions along the west coast of southern
Africa, where large concentrations of
gelatinous macroplankton occur (Brierley
et al. 2001). They indeed moved equator-
ward (Fig. 2), marginally following the
weak flow of the Benguela Current and
keeping outside the main corridor of in-
tense northwest-oriented Agulhas rings
(Schouten et al. 2000). They stayed outside
the intense part of the coastal upwelling
and only at a latitude of 26° S did they
move toward it. There is no clear indica-
tion from any environmental data that this
coastal approach was induced by any
ocean currents.
Mesoscale eddies
Eddies represent a relevant oceano-
graphic feature in the study region, and
indeed many of them were identified
during the turtle tracking period. In most
294
Fig. 4. Tracks of the turtle ‘Lara’ (thick line) and of a surface drifter
(no. 9421931; thin line) moving in the Indian Ocean in 1996. NS: nesting site
Lambardi et al.: Influence of ocean current on sea turtle movement
cases, the identified eddies were not ephemeral and
were moving either poleward (in the Indian Ocean), or
equatorward (in the Atlantic). The 8 turtles tracked in
oceanic waters encountered 42 eddies, for a total of
33 087 km (49.8% of total route lengths) covered
within, or in close association with, eddies.
The turtle routes were usually influenced to some
extent by the currents associated with the eddies. In 26
cases (61.9%), the encounter with the ro-
tating water masses within the eddy deter-
mined a deflection in the turtles’ move-
ment, evident as a curve or a bulge, which
was always in accordance with the current
sense of rotation. A clear example of this
pattern is illustrated by the looping move-
ments of ‘Alice’ and ‘Elena’. ‘Alice’ moved
directly away from the coast (Fig. 6) and
continued moving eastward with no evi-
dence whatsoever that she had crossed
into or across the Agulhas Current. The
positive SSHA present at the coast (red in
Fig. 6a), may well support such an infer-
ence by indicating the temporary absence
of the Agulhas Current itself. Indeed,
analysis of the appropriate SSHA images
showed that no such anomalies were pre-
sent offshore when the turtles that
moved SW left inshore waters. ‘Alice’ then
found a number of positive and negative
SSHA, which most probably were in fact
cyclonic and anti-cyclonic eddies, as
shown by their characteristic motion pole-
ward (Fig. 6) and by independent investi-
gations (Schouten et al. 2003). During the
passage of the first anti-cyclone (red in
Fig. 6a,b) the turtle carried out a half orbit
around this eddy (21 to 28 February). With
the advent of a persisting cyclonic anom-
aly (Fig. 6b,c), ‘Alice’ moved in a clock-
wise manner, encircling this eddy once
(1 March to 14 April), despite the fact that
the transmitter discontinuous duty cycle
prevented a detailed route reconstruction.
The turtle then made another shorter loop
that led her to encounter a strong anti-
cyclonic eddy (Fig. 6d). Her successive
movements were also completely in accor-
dance with the anti-clockwise water
motion associated with this large feature.
Although tracked in a different year,
‘Resi’ (Fig. 7) behaved consistently with
‘Alice’. ‘Resi’ also left the coast at a
right angle, but her movement was
then diverted anti-clockwise by a pass-
ing anti-cyclone (red in Fig. 7a; here
too, location data were received every 6 d). Her suc-
cessive northward path derived from her encounter
with poleward-moving eddies, as best illustrated by
her loop around (and in accordance with) a strong
anomaly farther north (Fig. 7b). ‘Resi’ finally
returned offshore of the nesting beach, at a time
when no significant anomaly was present, and then
started to move SW (Fig. 2) — a clear indication that
295
Fig. 5. (a) Route of the turtle ‘Sara’ as she entered the southern Atlantic Ocean
superimposed on a sea surface temperature (SST) image recorded on 15 July
2003. (b) Details of her inter-oceanic shift (dates shown) superimposed on a sea
surface height anomaly (SSHA) image recorded on 18 March 2003. Direction of
eddy rotation in the area revealed by SSHA indicated by dashed arrows: anti-
clockwise water movements are associated with anti-cyclonic eddies (positive
anomalies; red) and clockwise water movements with cyclonic eddies (negative
anomalies; blue). Black arrows indicate the turtle’s direction of movement
Mar Ecol Prog Ser 353: 289–301, 2008
she had been captured by the Agulhas Current
mainstream.
In other cases, the turtles remained for a long time
within the same eddy, with their routes correspond-
ingly showing 1 or more revolutions. This was ob-
served in 12 instances (28.6%), and the track sense of
rotation was always in accordance with that of the
eddy-associated currents. A good example is provided
by the prolonged circuitous movements of ‘Nemi’ and
‘Sandra’ discussed by Luschi et al. (2003b).
Finally, in 4 cases (9.5%) the turtles passed through
eddies with no discernible influence on their routes.
This happened especially with less intense anomalies
associated with weaker currents (but see Fig. 1b in
Luschi et al. 2003b for a crossing of an intense eddy).
For instance, during the northward movement offshore
of Namibia, ‘Sara’ encountered 2 eddies, but her tra-
jectory was only minimally influenced and she was
clearly able to move through them.
We also evaluated how long the turtles remained as-
sociated with the eddies they encountered (i.e. inde-
pendently from the kind of movements displayed dur-
ing these periods). In some cases, turtles remained
within the same persisting eddy for weeks (Luschi et al.
296
Fig. 6. Initial movements of the turtle ‘Alice’ upon leaving her nesting beach in 2001 (dates shown), superimposed on sea surface
height anomaly (SSHA) maps averaged for 4 successive periods, ending on (a) 28 February, (b) 25 March, (c) 14 April and (d) 10
May 2003. The white parts of the tracks correspond to the 10 d periods to which the SSHA images refer. The red dot indicates the
location of the nesting beach. Other explanations as in Fig. 5
Lambardi et al.: Influence of ocean current on sea turtle movement
2003b), and sometimes even followed them during their
poleward movement (see animation, available as Sup-
plementary Material at www.int-res.com/articles/
suppl/m353p289_app.gif). A similar behaviour was
shown by those turtles that, in the presence of moving
eddies, continuously shifted from 1 eddy to another
while remaining in the same geographical area, as de-
scribed above for ‘Resi’ and ‘Alice’.
A prolonged (>1 wk) uninterrupted
stay in eddies was observed in 8 in-
stances. In these cases, it is clear that
mesoscale features provided profitable
foraging opportunities to leatherbacks.
Signals of enhanced productivity in the
eddies were sometimes detectable in
ocean colour images. For instance, the
prolonged gyrations of ‘Nemi’ and ‘San-
dra’ occurred in an area with chl adensi-
ties >0.3 mg m–3, which coincides with a
semi-permanent northward bulge in
the Subtropical Convergence (Weeks &
Shillington 1996), exhibiting enhanced
concentrations of phytoplankton and of
higher trophic levels (Lutjeharms et al.
1986, Barange et al. 1998). On the other
hand, no enhanced biological activities
were observed in the area at about 34° S,
35° E where ‘Nemi’ circled for 3 mo.
DISCUSSION
Influence of currents on turtle routes
By matching satellite tracking data and
oceanographic information, we have
documented how pelagic-dwelling
leatherback turtles are substantially in-
fluenced by surface currents and related
features during their oceanic move-
ments. This influence takes place at 2
levels. (1) A large percentage of the
tracked turtle movements was per-
formed in total accordance with the cur-
rent flow, either when turtles were
moving along the Agulhas Current
mainstream or encountered rotating
water masses. On average, 73% of the
total tracked routes were determined by
currents. For the single turtles (Table 1),
this percentage was never less than
65%, the only exception being ‘Ronel’,
for which no clear relationship with
currents was established for her long
stay in front of the South African coast.
(2) Leatherback foraging movements were related to
specific oceanographic elements such as mesoscale
eddies, convergence or upwelling areas, which con-
centrate the macroplanktonic resources on which the
leatherbacks feed. The planktivorous turtles took
advantage of the profitable foraging opportunities
offered by these features, often remaining associated
with them for prolonged periods.
297
Fig. 7. Initial movements of the turtle ‘Resi’ in 2000 (dates shown) super-
imposed on sea suurfface height anomaly (SSHA) maps averaged for 2
successive periods: (a) 1 to 11 February and (b) 2 to 12 March 2003. Other
explanations as in Fig. 5
Mar Ecol Prog Ser 353: 289–301, 2008
The mechanical action of oceanic currents had pro-
found effects on the shape of the turtle routes and pro-
duced movement patterns as diverse as straight legs or
convoluted looping segments. In the first case, turtles
were transported quickly to distant regions, while in the
other situation they remained for weeks within the same
area. The influences of the currents on the turtle routes
often appeared substantial: for 7 turtles, long sections of
the routes were virtually indistinguishable from those of
inanimate drifters tracked in the same region. This sup-
ports previous conclusions (Luschi et al. 2003b) that
South African leatherbacks may passively drift with the
currents for long periods, although they can also move
independently of the currents. Gaspar et al. (2006) quan-
titatively demonstrated a similar prolonged drift in a
leatherback moving in the northern Atlantic Ocean. It is
therefore clear that attempting to infer leatherback
oceanic behaviour only by considering the reconstructed
(ground-related) routes may be misleading.
Under this scenario, extended parts of the
leatherback movements derive directly from current
advection rather than from a determined effort to swim
to specific distant sites, with the turtles possibly con-
centrating on diving to feed on planktonic prey. The
geographical displacements observed would then
essentially be a consequence of the fact that the
leatherback feeding areas are linked to major current
systems, and are spatio-temporally variable either over
short periods (i.e. months) or over the long term (i.e. in
different years). The turtle association with these mov-
able oceanic features would therefore account for the
wide extension of leatherback movements and for their
variability, for instance in separate sections of the same
route, in routes of turtles tracked in the same period or
in subsequent years. Investigation of the influence of
sea currents on leatherback movements has been lim-
ited to parts covered along oceanic fronts (Ferraroli et
al. 2004), and, especially, to a single turtle track in the
North Atlantic (Gaspar et al. 2006), where active trav-
elling coexisted with prolonged drifts with the current.
Occurrence of legs covered independently or against
currents and of segments influenced by oceanographic
features has been shown in pelagic-stage loggerhead
turtles Caretta caretta (Polovina et al. 2004, 2006).
Leatherback foraging strategies
We have also shown how leatherbacks nesting in
South Africa disperse widely, reaching disparate feed-
ing sites such as the waters north and east of South
Africa, the Subtropical Convergence, or the upwelling
region off SW Africa. This vast dispersion resembles
those known for post-nesting leatherbacks of other
populations, which have been recorded on a compara-
ble number of individuals (n = 8, Morreale et al. 1996;
n = 7, Eckert & Sarti 1997; n = 12, Ferraroli et al. 2004;
n = 9, Hays et al. 2004a,b). In these cases, the presence
of segments similar between turtles (especially soon
after leaving the nesting areas) coexists with variabil-
ity in the final destinations and with inter-annual vari-
ations in the routes followed. Such a pattern is also evi-
dent in leatherbacks tracked during other stages of
their life cycle (James et al. 2005).
In our study, most turtles reached foraging areas in
the oceanic environment, but in some cases foraging
occurred in coastal areas, even close to the nesting
beach (e.g. ‘Elena’). In particular, the deep-sea eddies
off the east coast of Africa represented a most relevant
oceanic feature for tracked leatherbacks, which
remained associated with the eddies for extended peri-
ods or even followed them during their poleward
movement. Most likely, eddies form patches of high
prey abundance, with turtles tending to remain within
them until these resources are totally exploited.
Mesoscale oceanographic features support well-struc-
tured food chains and are selectively targeted by
planktivorous animals like sharks, albatrosses and tur-
tles (Sims & Quayle 1998, Nel et al. 2001, Ferraroli et
al. 2004, Polovina et al. 2006). While the actual prof-
itability of eddies for leatherbacks may not be univer-
sal (Hays et al. 2006), in this region such features likely
offer valuable feeding opportunities to ocean-moving
leatherbacks, therefore constituting a kind of feeding
‘hotspot’ (Ferraroli et al. 2004, Hays et al. 2006, Polov-
ina et al. 2006), albeit movable, for planktivorous
leatherbacks.
One important characteristic of the eddies in this
area is their general high mobility, evident both in the
Indian and the Atlantic Ocean. As such, the pre-
dictability in time and space of these foraging hotspots
seems quite limited. This contrasts with other prof-
itable foraging areas of the region such as the Subtrop-
ical Convergence and the coastal upwelling region
west of the continent. Getting into these areas would
even be energetically favourable by riding the Agul-
has and Benguela Currents, but this was not the strat-
egy most commonly chosen by the tracked turtles, of
which only 3 joined the Convergence or the upwelling
zone. It may be somewhat surprising that South
African leatherbacks do not aim to reach such pre-
dictably profitable areas (at least in the first months
after nesting), but rather moved over extensive oceanic
areas. Even the rapid SW transfer carried along (and
with the help of) the Agulhas Current mainstream does
not seem to be due to the need to reach specific forag-
ing sites. One possible reason may be that within the
Subtropical Convergence, primary productivity is
extremely patchy in time and space (Llido et al. 2005),
so resource exploitation may not be particularly
298
Lambardi et al.: Influence of ocean current on sea turtle movement
advantageous, at least no more advantageous than in
lower-latitude areas, where food is less concentrated
but foraging is possibly equally efficient (e.g. within
eddies). No detailed information on the short-scale dis-
tribution of the gelatinous zooplankton is available
(James et al. 2005), and such subtle information can
hardly be derived from remote-sensing data. However,
it is worth noting that in some cases this SW leg was
postponed (‘Resi’) or absent for months (‘Alice’ and
‘Ronel’), indicating that leatherbacks do not have an
inherent or urgent need to move poleward.
This analysis suggests that South African leather-
backs do not aim to reach specific (e.g. previously
visited) destinations during their post-nesting move-
ments, but rather take advantage of whichever suit-
able oceanographic features they encounter during
their largely current-determined movements. Even for
those few turtles that have reached predictable fea-
tures, the arrival into these profitable areas was heav-
ily determined by the current conditions during the
journey (e.g. stochastic factors south of the continent),
and do not seem to derive from a turtle’s deliberate
movement toward one site or another. In the North
Atlantic, 2 foraging strategies seem to occur in post-
nesting leatherbacks, with some turtles heading to-
ward predictable feeding hotspots (like high-latitude
waters or eddy-rich regions; Ferraroli et al. 2004,
James et al. 2005), while others keep moving and con-
tinuously forage as they travel over large regions
(Hays et al. 2006). South African turtles seem to mainly
adopt this latter strategy, having the additional oppor-
tunity of encountering movable foraging hotspots such
as offshore eddies.
Implications for leatherback biology
The wide-ranging movements observed and espe-
cially the inter-oceanic shifts may play a major role in
determining the shallow genetic differences between
leatherbacks nesting in the Indian and Western
Atlantic Ocean (Dutton et al. 1999). However, since
these differences are related to maternal (mtDNA) lin-
eages, other processes such as nest site fidelity and
natal beach homing are probably involved in reducing
the observed genetic variability. Additionally, the fact
that South African leatherbacks do not necessarily
engage in long migrations may result in a reduction of
their remigration interval between nesting seasons.
However, remigration intervals of South African nest-
ing populations (typically 2 to 3 yr; Hughes 1996) are
similar to those of Atlantic populations (Boulon et al.
1996, Girondot et al. 2002), although somewhat shorter
than in Pacific leatherbacks (Spotila et al. 2000, Reina
et al. 2002).
Further, the strong influence of ocean currents on the
observed movements raises the issue of the naviga-
tional abilities of these turtles. South African leather-
backs return to the same area in successive nesting sea-
sons (Hughes 1996), and such fidelity requires reliance
on remarkable navigational abilities if leatherbacks
spend the 2 or 3 yr remigration intervals moving over
vast areas while being subjected to the current drifting
action. In the open sea, such drift cannot be detected,
and would take place both when turtles move in accor-
dance with the current flow (i.e. drift) and when they
are actively swimming toward specific destinations
(Girard et al. 2006). Drift-induced deviations from the
intended course are particularly harmful (and naviga-
tionally challenging) in the presence of strong and
varying currents, as in the present case. A possible help
to turtles in relocating the nesting area might come
from the existence of large oceanic gyres recirculating
most of the water in the South Indian Ocean (Stramma
& Lutjeharms 1997). The same current circulation that
leads the turtles away from the nesting area could
therefore carry them back to it (or close to it) over some
years. Such a system, however, would provide an im-
precise way of relocating the nesting beach and is not
available to turtles leaving the Indian Ocean gyre, e.g.
those entering the Atlantic Ocean. This scenario might
explain the irregularity in intervals between nesting
seasons exhibited by South African leatherbacks, but
could hardly explain their overall nesting beach fidelity
and the fact that no leatherback tagged in Tongaland
has ever been found nesting anywhere else in the
world (Hughes 1996). The best system by which ani-
mals subjected to drift can navigate back to a specific
site is to possess large-scale position-fixing mecha-
nisms. Reliance on such map-like mechanisms has
often been proposed for marine navigators including
turtles (Bingman & Cheng 2005), with experimental ev-
idence having been collected (Lohmann et al. 2004).
Such large-scale maps would help turtles subjected to
current drift to relocate the nesting beach, if only in
estimating how far from their target they have come
during their prolonged stay in a movable medium.
Acknowledgements. Funding for satellite tracking experi-
ments was provided by the Italian MiUR, the Italian Space
Agency and the UK Natural Environmental Research Council.
We are grateful to F. Papi, A. Sale and S. Benvenuti for their
help. Three anonymous referees provided useful comments
on an earlier version of the manuscript.
LITERATURE CITED
Alerstam T, Hedenström A, Åkesson S (2003) Long-distance
migration: evolution and determinants. Oikos 103:247–260
Barange M, Pakhomov EA, Perissinotto R, Froneman PW,
Verheye HM, Taunton-Clark J, Lucas MI (1998) Pelagic
community structure of the Subtropical Convergence
299
Mar Ecol Prog Ser 353: 289–301, 2008
region south of Africa and in the mid-Atlantic Ocean.
Deep-Sea Res 45:1663–1687
Beal LM, Bryden HL (1999) The velocity and vorticity struc-
ture of the Agulhas Current at 32° S. J Geophys Res 104:
5151–5176
Bingman VP, Cheng K (2005) Mechanisms of animal global
navigation: comparative perspectives and enduring chal-
lenges. Ethol Ecol Evol 17:295– 318
Boulon RH Jr, Dutton PH, McDonald DL (1996) Leatherback
turtles (Dermochelys coriacea) on St. Croix, U.S. Virgin
Islands: fifteen years of conservation. Chelonian Conserv
Biol 2:141–147
Brierley AS, Axelsen BE, Beucher E, Sparks CAJ, Boyer H,
Gibbons MJ (2001) Acoustic observations of jellyfish in the
Namibian Benguela. Mar Ecol Prog Ser 210:55–66
de Ruijter WPM, Ridderinkhof H, Lutjeharms JRE, Schouten
MW, Veth C (2002) Observations of the flow in the
Mozambique Channel. Geophys Res Lett 29:1401–1403
de Ruijter WPM, van Aken HM, Beier EJ, Lutjeharms JRE,
Matano RP, Schouten MW (2003) Eddies and dipoles
around South Madagascar: formation, pathways and
large-scale impact. Deep-Sea Res Part I 51(3):383–400
Dingle H (1996) Migration, the biology of life on the move.
Oxford University Press, New York
Duron-Dufrenne M (1987) Premier suivi par satellite en
Atlantique d’une tortue Luth Dermochelys coriacea. CR
Acad Sci Paris 304:399– 402
Dutton PH, Bowen BW, Owens DW, Barragan A, Davis SK
(1999) Global phylogeography of the leatherback turtle
(Dermochelys coriacea). J Zool (Lond) 248:397– 409
Eckert SA (2002) Swim speed and movement patterns of
gravid leatherback sea turtles (Dermochelys coriacea) at
St Croix, US Virgin Islands. J Exp Biol 205:3689–3697
Eckert SA (2006) High-use oceanic areas for Atlantic
leatherback sea turtles (Dermochelys coriacea) as identi-
fied using satellite telemetered location and dive informa-
tion. Mar Biol 149:1257–1267
Eckert SA, Sarti L (1997) Distant fisheries implicated in the
loss of the world’s largest leatherback nesting population.
Mar Turtle Newsl 78:2–7
Ferraroli S, Georges JY, Gaspar P, Le Maho Y (2004) Where
leatherback turtles meet fisheries. Nature 429:521–522
Flemming B, Hay R (1988) Sediment distribution and dynam-
ics of the Natal continental shelf. In: Schumann EH (ed)
Coastal ocean studies off Natal, South Africa. Springer-
Verlag, Berlin, p 47–80
Gaspar P, George JY, Fossette S, Lenoble A, Ferraroli S, Le
Maho Y (2006) Marine animal behaviour: neglecting
ocean currents can lead us up the wrong track. Proc R Soc
Lond B 237:2697–2702
Girard C, Sudre J, Benhamou S, Roos D, Luschi P (2006) Hom-
ing in green turtles Chelonia mydas: oceanic currents act
as a constraint rather than as an information source. Mar
Ecol Prog Ser 322:281–289
Girondot M, Tucker AD, Rivalan P, Godfrey MH, Chevalier J
(2002) Density-dependent nest destruction and population
fluctuations of Guianan leatherback turtles. Anim Conserv
5:75– 84
Gründlingh ML (1983) On the course of the Agulhas Current.
S Afr Geogr J 65:49– 57
Hays GC, Houghton JDR, Isaacs C, King RS, Lloyd C, Lovell P
(2004a) First records of oceanic dive profiles for leather-
back turtles (Dermochelys coriacea) indicate behavioural
plasticity associated with long distance migration. Anim
Behav 67:733– 743
Hays GC, Houghton JDR, Myers AE (2004b) Pan-Atlantic
leatherback turtle movements. Nature 429:522
Hays GC, Hobson VJ, Metcalfe JD, Righton D, Sims DW
(2006) Flexible foraging movements of leatherback turtles
across the north Atlantic Ocean. Ecology 87:2647–2656
Hughes GR (1996) Nesting of the leatherback turtle (Der-
mochelys coriacea) in Tongaland, KwaZulu-Natal, South
Africa, 1963–1995. Chelonian Conserv Biol 2:153–158
James MC, Myers RA, Ottensmeyer CA (2005) Behaviour of
leatherback sea turtles, Dermochelys coriacea, during the
migratory cycle. Proc R Soc Lond B 272:1547–1555
Keinath JA, Musick JA (1993) Movements and diving behav-
ior of a leatherback turtle, Dermochelys coriacea. Copeia
1993:1010–1017
Llido J, Garçon V, Lutjeharms JRE, Sudre J (2005) Event-scale
blooms drive enhanced primary productivity at the Sub-
tropical Convergence. Geophys Res Lett 32:L15611
Lohmann KJ, Lohmann CMF, Ehrhart LM, Bagley DA, Swing
T (2004) Geomagnetic map used in sea-turtle navigation.
Nature 428:909–910
Luschi P, Hays CG, Papi F (2003a) A review of long-distance
movements by marine turtles and the possible role of
ocean currents. Oikos 103:293–301
Luschi P, Sale A, Mencacci R, Hughes GR, Lutjeharms JRE,
Papi F (2003b) Current transport in leatherback sea turtles
(Dermochelys coriacea) in the ocean. Proc R Soc Lond B
270 (Suppl 2):129–132
Lutcavage ME, Rhodin AGJ, Sadove SS, Conroy CR (2002)
Direct carapacial attachment of satellite tags using ortho-
pedic bioabsorbable mini-anchor screws on leatherback
turtles in Culebra, Puerto Rico. Mar Turtle Newsl 95:9–12
Lutjeharms JRE (2006) The Agulhas Current. Springer-Verlag,
Berlin
Lutjeharms JRE, Ansorge I (2001) The Agulhas Return Cur-
rent. J Mar Syst 30:115–138
Lutjeharms JRE, van Ballegooyen RC (1988) Anomalous
upstream retroflection in the Agulhas Current. Science
240(4860):1770–1772
Lutjeharms JRE, Allanson BR, Parker L (1986) Frontal zones,
chlorophyll and primary production patterns in the sur-
face waters of the Southern Ocean south of Africa. In:
Nihoul JCJ (ed) Marine interfaces ecohydrodynamics.
Proc 17th Int Liege Colloquium. Elsevier Oceanography
Series, p 105–117
Lutjeharms JRE, Shillington FA, Duncombe Rae CM (1991)
Observations of extreme upwelling filaments in the South
East Atlantic Ocean. Science 253(5021):774– 776
Lutjeharms JRE, Weeks SJ, van Ballegooyen RD, Shillington
FA (1992) Shedding of an eddy from the seaward front of
the Agulhas Current. S Afr J Sci 88:430–433
Morreale SJ, Standora EA, Spotila JR, Paladino FV (1996)
Migration corridor for sea turtles. Nature 384:319 –320
Musick JA, Limpus CJ (1997) Habitat utilization and migration
in juvenile sea turtles. In: Lutz PL, Musick JA (eds) The biol-
ogy of sea turtles. CRC Press, Boca Raton, FL, p 137–163
Nel DC, Lutjeharms JRE, Pakhomov EA, Ansorge IJ, Ryan
PG, Klages NTW (2001) Exploitation of mesoscale oceano-
graphic features by grey-headed albatross Thalassarche
chrysostoma in the southern Indian Ocean. Mar Ecol Prog
Ser 217:15–26
Polovina JJ, Balazs GH, Howell EA, Parker DM, Seki MP,
Dutton PH (2004) Forage and migration habitat of logger-
head (Caretta caretta) and olive ridley (Lepidochelys oli-
vacea) sea turtles in the central North Pacific Ocean. Fish
Oceanogr 13:36– 51
Polovina J, Uchida I, Balazs GH, Howell EA, Parker D, Dutton
P (2006) The Kuroshio Extension Bifurcation Region: a
pelagic hotspot for juvenile loggerhead sea turtles. Deep-
Sea Res Part II 53:326–339
300
Lambardi et al.: Influence of ocean current on sea turtle movement
Reina RD, Mayor PA, Spotila JR, Piedra R, Paladino FV (2002)
Nesting ecology of the leatherback turtle, Dermochelys
coriacea, at Parque Nacional Marino Las Baulas, Costa
Rica: 1988– 89 to 1999–2000. Copeia 2002:653–664
Richardson PL, Lutjeharms JRE, Boebel O (2003) Introduction
to the ‘Inter-ocean exchange around southern Africa’.
Deep-Sea Res Part II 50:1–12
Sale A, Luschi P, Mencacci R, Lambardi P, Hughes GR, Hays
GC (2006) Long-term monitoring of leatherback turtle div-
ing behaviour during oceanic movements. J Exp Mar Biol
Ecol 328:197–201
Schouten MW, de Ruijter WPM, van Leeuwen PJ, Lutjeharms
JRE (2000) Translation, decay and splitting of Agulhas
rings in the south-eastern Atlantic ocean. J Geophys Res
105:21913–21925
Schouten MW, de Ruijter WPM, van Leeuwen PJ, Rid-
derinkhof H (2003) Eddies and variability in the Mozam-
bique Channel. Deep-Sea Res Part II 50:1987–2003
Shannon LV, Nelson G (1996) The Benguela: large scale fea-
tures and processes and system variability. In: Wefer G,
Berger WH, Siedler G, Webb D (eds) The South Atlantic:
present and past circulation. Springer-Verlag, Berlin,
p 163–210
Sims DW, Quayle VA (1998) Selective foraging behaviour of
basking sharks on zooplankton in a small-scale front.
Nature 393:460–464
Spotila JR, Reina RD, Steyermark AC, Plotkin PT, Paladino
FV (2000) Pacific leatherback turtles face extinction.
Nature 405:529–530
Stewart RH (2003) Introduction to physical oceanography.
Available at http://oceanworld.tamu.edu/resources/ocng_
textbook/contents.html
Stramma L, Lutjeharms JRE (1997) The flow field of the sub-
tropical gyre of the South Indian Ocean. J Geophys Res
102:5513– 5530
Wedepohl PM, Lutjeharms JRE, Meeuwis JM (2000) The sur-
face drift of the South East Atlantic Ocean. S Afr J Mar Sci
22:71–79
Weeks SJ, Shillington FA (1996) Phytoplankton pigment
distribution and frontal structure in the subtropical
convergence region south of Africa. Deep-Sea Res 43:
739– 768
301
Editorial responsibility: Rory Wilson (Contributing Editor),
Swansea, UK
Submitted: June 23, 2006; Accepted: June 21, 2007
Proofs received from author(s): December 7, 2007