BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1):7-16, 2009

APPLICATION OF REMOTE SENSING TO THE STUDY OF THE PELAGIC SPINY

LOBSTER LARVAL TRANSPORT IN THE TROPICAL ATLANTIC*

Camila Aguirre Góes Rudorff¹; João Antônio Lorenzzetti¹; Douglas F. M. Gherardi¹

and Jorge Eduardo Lins-Oliveira²
1
Instituto Nacional de Pesquisas Espaciais - INPE.
(Av. dos Astronautas, 1758, 12245-970 São José dos Campos, SP, Brasil)
camila.rudorff@gmail.com, loren@dsr.inpe.br, douglas@dsr.inpe.br

²Universidade Federal do Rio Grande do Norte - UFRN/DOL

(Av. Praia de Mãe Luíza s/n, 59014-100 Natal, RN, Brasil)
jorgelins@ufrnet.br

ABSTRACT
The connectivity of marine populations via larval dispersal is crucial for the maintenance of fisheries
production and biodiversity. Because larval dispersion takes place on different spatial scales, global
operational satellite data can be successfully used to investigate the connectivity of marine
populations on different spatial and temporal scales. In fact, satellite data have long been used for the
study of the large and mesoscale biological processes associated with ocean dynamics. This paper
presents simulations of spiny lobster larvae transport in the Tropical Atlantic using the geostrophic
currents, generated by altimetry that feeds an advection/diffusion model. Simulations were
conducted over the Tropical Atlantic (20oN to 15oS), considering four larvae release areas: the Cape
Verde Archipelago, the Ivory Coast, Ascension Island and Fernando de Noronha Archipelago. We
used mean geostrophic current (MGC) calculated from 2001 to 2005 to represent the mean
circulation of the Tropical Atlantic. We also ran the model for the El Niño geostrophic current regime
(ENGC) using part of the MGC data, representing the El Niño 2002/2003 event. Results suggest that
the intensification of the mesoscale ocean processes associated with El Niño events promotes the
connectivity between populations, increasing the chances of a genetic flux among different stocks.
We concluded that the altimetry geostrophic current data together with a relatively simple
advection/diffusion model can provide useful information about the physical dynamics necessary to
conduct studies on larval dispersion.

RESUMO
A conectividade de populações marinhas através da dispersão larval é crucial para a manutenção da
produção pesqueira e da biodiversidade. A dispersão de larvas ocorre em diferentes escalas espaciais
e temporais, de forma que o recobrimento global e escala sinóptica fazem dos dados de satélite
ferramentas importantes para esses estudos. O objetivo deste artigo é apresentar os resultados do uso
de dados de correntes geostróficas derivadas de satélites altímetros para simular o transporte de
larvas de lagosta espinhosa no oceano Atlântico Tropical. As simulações foram realizadas ao longo
do Atlântico Tropical (20oN - 15oS), iniciando em quatro locais (Cabo Verde, Costa do Marfim, Ilha
de Ascensão e Arquipélago de Fernando de Noronha). Foi utilizado um modelo advectivo/difusivo
forçado com o campo médio de circulação geostrófica calculado entre 2001 e 2005 e outro forçado
com campo correspondente ao evento de El Niño 2002/2003. Os resultados obtidos sugerem que há
uma intensificação de processos oceânicos de meso-escala durante o ano de El Niño, que promove a
conectividade entre diferentes estoques e aumenta as chances de ocorrer um fluxo genético. Os dados
de correntes geostróficas gerados a partir de satélites altímetros incorporados a um modelo
advectivo/difusivo simples podem prover informações importantes acerca da dinâmica física
necessárias para conduzir estudos sobre dispersão larval.

Descriptors: Satellite data, Altimetry, Larval transport, El Niño.

Descritores: Dados de satélite, Altimetria, Transporte de larvas, El Niño.

__________
(*) Paper presented at 1st Seminary on Remote Sensing Applied to Fishing, on 11-12 September 2006. São José dos Campos,
SP, Brazil.
8 BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1), 2009

INTRODUCTION Modeling Early Life Stages

The majority of marine organisms exhibit a

complex life cycle that includes separate planktonic
Many marine benthic species have
larval and bottom-dwelling juvenile and adult phases.
populations connected by larval transport during their
Dispersal in its broadest sense means movement away
planktonic phases. Connectivity depends largely on
from the birthplace (SUGDEN; PENNISI, 2006).
the larval duration in the plankton, the spatial
Besides the physical conditions affecting larval
heterogeneities in physical conditions such as
dispersion, biological parameters are extremely
advection and diffusion, vertical larval migration, and
important. Among the most important biological
mortality (COWEN et al., 2000; BECKER et al.,
parameters, two deserve special attention: the time of
2007). These are important elements in fisheries
development and the swimming behavior of larvae.
management and in the design of marine protected
The first one should be divided into two categories:
areas (MPAs) that have fuelled the debate over the
short-lived planktonic larvae (e.g. reef fish larvae and
relative importance of size and spacing of MPAs
other invertebrate larvae) and long-lived planktonic
(HASTINGS; BOTSFORD, 2006; ALMANY et al.,
larvae (e.g. spiny lobster larvae). Short-lived
2007). Connectivity among spatially structured marine
planktonic larvae, especially fish larvae, seem to
populations is also known to be important to support
present a self-recruitment process more intense and
sustainable fisheries.
efficient than previously thought (COWEN et al.,
Because larval dispersion takes place on
2006; ALMANY et al., 2007).
different spatial scales, it has become increasingly
Fish larvae present intense swimming
evident that satellite data can be successfully used to
abilities; therefore models should better address
investigate the connectivity of marine populations on
mesoscale features and, if possible, simulate the water
different spatial and temporal scales. For example,
column’s behavior (LEIS, 2006; MONTGOMERY et
since the mid-seventies, maps of sea surface
al., 2006). In contrast, the pelagic larvae of spiny
temperature (SST) have been derived from infrared
lobsters (called phyllosoma) show some adaptations
radiometers onboard NOAA (National Oceanic and
inherent to holoplankton (i.e. a transparent and
Atmospheric Administration) satellites. Based on
dorsoventrally flattened leaf-like body) and are
these SST maps it is possible, for instance, to extract
associated with their ability to withstand up to 12
information about the ocean currents, their frontal
months in the plankton. These characteristics and the
zones, meanderings and eddy shedding. Long-term
lack of swimming abilities allow spiny lobster larvae
global SST data sets derived from satellites are also
to be transported for long distances by ocean currents
considered one of the most useful indicators of climate
away from the spawning area (PHILLIPS et al., 1980;
change. Ocean color sensors have been used to
PHILLIPS; SASTRY, 1980; BOOTH; PHILLIPS,
monitor variations of chlorophyll-a concentrations on
1994).
a global scale, which are of great importance for
Studies on lobster larval dispersion using
biological studies of marine ecosystems and to climate
geostrophic currents derived from altimetry data
change, considering that the CO2 oceanic intake is
include that of Chiswell and Booth (1999) near the
controlled by the planktonic photosynthesis.
Wairarapa eddy (New Zealand), using T/P satellite
Altimetric radar such as TOPEX/Poseidon (T/P),
data. Polovina et al. (1999) also used T/P data to force
European Remote-Sensing satellite (ERS) and Jason
an advection/diffusion model for P. marginatus in the
provide information to produce maps of dynamic
Hawaiian Archipelago. Griffin et al. (2001) developed
height on an operational and global basis, and derive
a quantitative biophysical model of lobster larval (P.
geostrophic current maps with a temporal resolution of
cygnus) transport using a T/P geostrophic velocities
7-10 days. These maps can be used to force
field. More recently, Chiswell et al. (2003) used
advection/diffusion models of marine larvae,
geostrophic current fields derived from altimetry to
particularly for spiny lobster larvae that can be
run a larval transport model in the Tasman Sea.
advected for long periods by ocean currents.
In the Tropical Atlantic Ocean there are five
It is possible to simulate larval advection
spiny lobster species of the genus Panulirus, of which
using the current field generated as the output of ocean
P. argus, P. leavicauda and P. echinatus occur in
models. These rather complex models are normally
Brazilian waters, including the oceanic islands of the
initialized with climatological fields of temperature
Atol das Rocas, the Fernando de Noronha and the São
and salinity and are forced by either climatological or
Pedro São Paulo archipelago (HOLTHUIS, 1991).
satellite winds. However, geostrophic current fields
There is a predominance of P. echinatus in the
derived from spaceborne altimetry data can be used to
oceanic islands of Ascension, São Pedro and São
force much simpler advection/diffusion models.
Paulo, Fernando de Noronha and the Atol das Rocas
 RUDORFF ET AL.: APPLICATION OF REMOTE SENSING TO THE STUDY OF THE PELAGIC LARVAL TRANSPORT 9

(COELHO; RAMOS-PORTO, 1998; TAVARES, NOAA to characterize El Niño events. This Index
2003). The species P. echinatus, P. regius and P. refers to the sea surface temperature anomaly (SSTA)
charlestoni are known to occur off the northwestern taken in the Tropical Pacific (5oN-5oS and 120o-
African coast. However, Freitas and Castro (2005) 170oW) based on the climatology from 1971 to 2000.
reported the occurrence of P. argus in the Cape Verde According to NOAA, El Niño and La Niña periods are
Archipelago and also cited other occurrences off the identified in the series when a 0.5oC negative (El
Ivory Coast (see also Tavares, 2003). Niño) or positive (La Niña) anomaly occurs during 5
Grimm (1999) suggests a few basic rules consecutive months. According to this Index we
when dealing with ecological models: (i) models selected the geostrophic current velocities database
should be kept as simple as possible; (ii) the final from April 3rd 2002 to April 2nd 2003 time-series to
importance of the model is to understand the processes represent the El Niño phenomenon.
involved, and (iii) modelers should adopt an We correlated the u (r = 0.93, p=0.05, n =
experimental attitude. 70) and v (r = 0.63, p=0.05, n = 51) velocity
In the present paper we show how components from MGC with coincident velocities
geostrophic currents generated by altimetry can be obtained by drifting-buoys (Global Drifter Program -
used to simulate the transport of spiny lobster larvae, GDP/SVP, http://www.meds-sdmm.dfo-mpo.gc.ca/)
addressing the implications for the connectivity of for the study area (excluding buoys without drogue).
potential spawning grounds. We also evaluate the With the same GDP/SVP drifter database we
influence of El Niño in changing the velocity fields calculated the zonal (kx) and meridional (ky) eddy-
and the connectivity among these spawning grounds. diffusion coefficients, using only buoys with the
drogue attached (n=377), in accordance with the
MATERIALS AND METHODS methodology proposed by Assireu (2003). We
calculated kx and ky for 35 boxes of 2ox2o distributed
Ocean Current Data over the Tropical Atlantic, however, to represent the
turbulence in the study area, our model used mean
We used the geostrophic surface current values for kx (3.97x107cm²/s) and ky (2.05x107cm²/s).
velocity fields obtained from altimetry, distributed by
AVISO (http://www.aviso.oceanobs.com/), to run the Advection/Diffusion Model
advection/diffusion model. This database is a result of
merged observations from T/P, Jason and ERS The advection/diffusion model was run by
altimeters to achieve improved spatial and temporal iteratively applying successive advective
resolutions of 1/3 degree and 7 days in a global displacements due to water flow with added random
coverage. Poleward of 5ºN and 5ºS the zonal (u) and displacement (ε) associated with diffusion.
meridional (v) geostrophic velocity components are Starting at a chosen xt and yt initial position,
calculated from the meridional and zonal slopes of the the new location (xt+∆t, yt+∆t) of each virtual larvae
dynamic height (ζ) derived from the sea level after a time interval ∆t, is updated in time by the
anomalies (SLA). In the ±5º equatorial band, where equations 1 and 2 (POLOVINA et al., 1999).
the geostrophic assumption becomes weak, velocities
are calculated using the second derivative of ζ [u ( xt , yt ,t ) ∆t + ε kx∆t ]
according to Picaut (1989) and Lagerloef et al. (1999). xt + ∆ t = x t + (1)
The absolute dynamic topography used in the cos( yt )
calculations of u and v are obtained by adding SLA to
the mean dynamic topography (MDT), the latter being
estimated by subtracting the standard geoid from the yt + ∆ t = yt + [v( xt , yt ,t ) ∆t + ε ky∆t ] (2)
mean sea surface height (MSH).
We calculated the mean values of the zonal
(u) and meridional (v) components using the AVISO where:
database from 2001 to 2005 for each week of the year
(1 to 52), obtaining the mean geostrophic current t = time (day);
(MGC). This was the velocity field used to advect the x and y = position of the larvae (degrees of longitude
particles, representing the surface circulation of the and latitude);
Tropical Atlantic for a typical year. We also ran the u and v = zonal and meridional geostrophic velocity
model for the El Niño geostrophic current regime components (degree/day);
(ENGC) using part of the MGC data relative to the El ε = normal distributed random variable (zero mean,
Niño event that occurred between 2002 and 2003. The unit standard deviation);
El Niño period was determined using the Oceanic kx and ky = zonal and meridional eddy-diffusion
Niño Index (ONI), one of the standard indices used by coefficients (degree2/day).
 10 BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1), 2009

The first and second terms inside the termed only larvae) from randomly chosen points
brackets correspond to the advective and the diffusive inside a 1olat x 1olong rectangle. Larvae were advected
displacements, respectively. The cosine function in the for 365 days using mean geostrophic current (MGC)
first equation corrects for the poleward convergence of and the El Niño geostrophic current regime (ENGC).
meridians with latitude.
The time step was set to one day (∆t=1) with RESULTS
365 iterations representing one year of simulation.
Larvae have a new position each day, calculated using Plots of larval trajectories readily allow
the u and v geostrophic velocities linearly interpolated differentiation between the advective fields from the
from the four closest grid points. The simulation is MGC and the ENGC. The only exception is the Ivory
halted when the larvae hit the coastline or whenever Coast where simulations show almost no differences
any of the four grid interpolating points touches land. (Fig. 2).
Forcing altimeter geostrophic velocities are updated All other larval drifts from the El Niño
every seven days and no daily interpolation of this simulations present wider dispersal ranges when
data set was performed between successive weeks. compared to MGC runs (Fig. 3-5). Considering the
spawning ground of the Cape Verde Archipelago (Fig.
Simulations
3), when MGC data are used, dispersal boundaries are
somewhat confined for both spawning periods (April
Simulations were conducted over the
and September) and meridional larval transport did not
Tropical Atlantic (20oN to 15oS) starting during the
reach 5º N. The ENGC simulations, in contrast,
spawning peak seasons of April and September
indicate that larvae can drift as far south as the equator
(CAVALCANTE-SOARES; FONTELES-FILHO,
when spawning occurs in September. Trajectories
2000) from four different sites known to have adult
simulated for April have comparable boundaries but
populations: Cape Verde, Ivory Coast, Ascension
the ENGC dispersion is less compact during the El
Island and Fernando de Noronha Archipelago (Fig. 1).
Niño period.
Each model run released 5000 virtual larvae (hereafter

Fig. 1. Tropical Atlantic and study area (dashed rectangle). Main superficial ocean currents are indicated by arrows, including
the North Equatorial Current NEC, North Equatorial Countercurrent NECC, Guinea Current, Angola Current, northern, central
and southern branches of the South Equatorial Current SEC, Brazil Current, Guyana Current, North Brazil Current NBC and its
retroflection (NBC Retr.) (Adapted from Lumpkin and Garzoli, 2005). The stars ( ) mark the release areas used to start the
simulations: 1 - Cape Verde (24-25ºW, 15-16ºN); 2 - Ivory Coast (7-8ºW, 3-4ºN); 3 - Ascension Island (14-15ºW, 7-8ºS); 4 -
Fernando de Noronha Archipelago (31-32ºW, 3-4ºS).
 RUDORFF ET AL.: APPLICATION OF REMOTE SENSING TO THE STUDY OF THE PELAGIC LARVAL TRANSPORT 11

Abril
Ivory Coast Ivory Coast
15oN 15oN
MGC (April) ENGC (April)
A B

5oN 5oN

5oS 5oS

15oS o
15oS o 45 W 35oW 25oW 15oW 5oW 5oE 15oE
45 W 35oW 25oW 15oW 5oW 5oE 15oE

Ivory Coast
Ivory Coast
Setembro 15oN
o
15 N ENGC (Sep.)
MGC (Sep.)
D
C

5oN
5oN

5oS 5oS

15oS o 15oS o
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45 W 35oW 25oW 15oW 5oW 5oE 15oE

Fig. 2. Trajectories of virtual larvae (n= 5000) based on simulations starting in April (top) and September (bottom)
from Ivory Coast ( ). Right-hand maps represent El Niño geostrophic currents (ENGC) and left-hand maps represent
mean geostrophic currents (MGC). Circles indicate other adult stock populations to help visualize the connectivity
among sites. Larvae trajectories are colored to improve visualization.

Cape Verde Cape Verde

15oN
Abril 15oN
MGC (April) ENGC (April)
A B

5oN 5oN

5oS 5oS

15oS o o
45 W 35oW 25oW 15oW 5oW 5oE 15oE 15 45
So
W 35oW 25oW 15oW 5oW 5oE 15oE

Setembro
Cape Verde Cape Verde
o
115 N 15oN
MGC (Sep.) ENGC (Sep.)
C D

5oN 5oN

5oS 5oS

115oS o 15oS
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45oW 35oW 25oW 15oW 5oW 5oE 15oE

Fig. 3. Trajectories of virtual larvae (n= 5000) based on simulations starting in April (top) and September (bottom)
from Cape Verde ( ). Right-hand maps represent El Niño geostrophic currents (ENGC) and left-hand maps represent
mean geostrophic currents (MGC). Circles indicate other adult stock populations to help visualize the connectivity
among sites. Larvae trajectories are colored to improve visualization.
 12 BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1), 2009

Larval drifts originating from Ascension The overall intensification of the magnitudes
Island have more impressive differences between of ENGCs relative to MGCs for the whole Tropical
spawning seasons and geostrophic current data. The Atlantic is shown in Table 1. Except for the austral
MGC-based simulations display a more compact summer period, maximum weekly average velocities
trajectory pattern than the ENGC simulations, the are higher during the El Niño period than during the
latter extending zonally further to the east and rest of the year. Maximum average velocity was
meridionally to the north and south (Fig. 4). These particularly high during the El Niño winter.
north-south extensions of trajectories extrapolate for The spatial and seasonal variations in the
just a few degrees but what seems more relevant is the ENGC magnitudes of the geostrophic current
lack of compactness observed in the trajectories anomalies (relative to MGC) for the El Niño period
generated with the ENGC data. The simulation of were calculated at each grid point for the Tropical
trajectories departing from Ascension Island using the Atlantic. The results are displayed in Figure 6
MGC data starting in September (Fig. 4C) is the only superimposed on the mean geostrophic current vectors
one that shows no extensions towards the Brazilian calculated using the MGC database. Anomalies were
coast beyond 5ºS. calculated based on the resultant of the two
Spawning simulations dispersing from the geostrophic velocity components (u and v), so that
Fernando de Noronha Archipelago also show a distinct anomalies represent changes in magnitude (but not in
behavior during the El Niño period (Fig. 5). In this direction). We can observe the MGC flow direction
release area larvae trajectories display a wider and intensity by looking at the vectors overlying these
dispersion as compared to the MGC scenario, reaching anomalies graphs. An increase in the magnitude of
more eastern longitudes but also being found closer to geostrophic currents is most evident in the North
the northern Brazilian NE coast. It is clear that surface Brazil Current retroflection region during austral
advection around the Fernando de Noronha spring and winter and a weakening during the austral
Archipelago has a comparatively weaker meridional summer. Of greater importance, however, are the
component because most simulated trajectories have a positive and negative anomalies seen throughout the
predominant east-west orientation. An example of this equatorial Atlantic during the austral autumn and
zonal advection is the ENGC scenario starting in winter seasons, respectively.
September, by which some larvae were advected as far
as 22ºW (Fig. 5D).

Ascension Ascension
Abril 15oN
15oN
MGC (April) ENGC (April)
A B

5oN
5oN

5oS 5oS

15oS o 15oS o
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45 W 35oW 25oW 15oW 5oW 5oE 15oE

Ascension
Setembro Ascension
o 15oN
15 N
MGC (Sep.) ENGC (Sep.)
C D

5oN 5oN

5oS 5oS

15oS o 15oS o
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45 W 35oW 25oW 15oW 5oW 5oE 15oE

Fig. 4. Trajectories of virtual larvae (n= 5000) based on simulations starting in April (top) and September (bottom)
from Ascension ( ). Right-hand maps represent El Niño geostrophic currents (ENGC) and left-hand maps represent
mean geostrophic currents (MGC). Circles indicate other adult stock populations to help visualize the connectivity
among sites. Larvae trajectories are colored to improve visualization.
 RUDORFF ET AL.: APPLICATION OF REMOTE SENSING TO THE STUDY OF THE PELAGIC LARVAL TRANSPORT 13

Abril
Fernando de Noronha Fernando de Noronha
15oN 15oN
MGC (April) ENGC (April)
A B

5oN 5oN

5oS 5oS

15oS o 15oS o
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45 W 35oW 25oW 15oW 5oW 5oE 15oE

FernandoSetembro
de Noronha Fernando de Noronha
15oN 15oN
MGC (Sep.) ENGC (Sep.)
C D

5oN 5oN

5oS 5oS

15oS o 15oS
45 W 35oW 25oW 15oW 5oW 5oE 15oE 45oW 35oW 25oW 15oW 5oW 5oE 15oE

Fig. 5. Trajectories of virtual larvae (n= 5000) based on simulations starting in April (top) and September (bottom)
from Fernando de Noronha ( ). Right-hand maps represent El Niño geostrophic currents (ENGC) and left-hand
maps represent mean geostrophic currents (MGC). Circles indicate other adult stock populations to help visualize the
connectivity among sites. Larvae trajectories are colored to improve visualization.

Table 1. Maximum geostrophic velocities for the Tropical Atlantic from

SSALTO/DUACS data base (MGC) and data for the El Niño period.

Mean Geostrophic El Niño

Current (MGC) (ENGC)
Maximum Maximum
Summer 64.76 41.52
Autumn 38.93 54.33
Winter 45.02 91.28
Spring 58.13 63.93

The observed geostrophic velocity anomalies bands in the equatorial Atlantic. As a consequence, the
during an El Niño year suggest large spatial and connectivity of adult lobster populations via larval
seasonal variabilities. This indicates that the greatest dispersion in the Tropical Atlantic operates within a
impact of an El Niño event upon larval dispersion will complex system. For example, the Cape Verde
depend on the relative positions of the larvae cloud Archipelago located in the Tropical Eastern Atlantic,
and the regions of maximum geostrophic velocity is subject to different oceanographic regimes
anomalies. (STRAMMA et al., 2005), including major large scale
surface flow fields of the eastward North Equatorial
DISCUSSION Counter Current (NECC) and the Cape Verde frontal
zone (CVFZ) (ZENK et al., 1991) between the North
There is a permanent tropical circulation Equatorial Current (NEC) and the northern branch of
composed of several zonal current and countercurrent the North Equatorial Counter Current (nNECC). It is
also close to mesoscale variabilities such as tropical
14 BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1), 2009

Austral Summer

Austral Autumn

Austral Spring

Austral Winter

cm/s
Fig. 6. Magnitudes of the seasonal geostrophic current anomalies (scale in colors) for the El Niño
period calculated from the four year mean (MGC). These magnitudes are superimposed by the
four year mean geostrophic velocity vectors (black arrows).
 RUDORFF ET AL.: APPLICATION OF REMOTE SENSING TO THE STUDY OF THE PELAGIC LARVAL TRANSPORT 15

instability vortices and zonal fronts. These model outputs. On the other hand, it also shows that
perturbations originate from mid-May to September in regional instabilities have to be considered in any
the central and eastern basin translating westward with analysis of either short or long-lived larval dispersion
a maximum speed of 40 cm s-1 (FOLTZ et al., 2004). of marine organisms as these may contribute to larval
Our simulations of larval dispersal from Cape Verde retention or long-distance dispersion. We contend that
capture this dynamically complex scenario as showed an unbiased interpretation of larval trajectories and the
by wider dispersion trajectories beginning in April connectivity they promote must necessarily be based
(Fig. 3A). During the El Niño period, these mesoscales on a sound understanding of the dynamics of surface
features are intensified and larvae can be advected ocean circulation and their mesoscale turbulence
further southeast, connecting lobster stocks from Cape processes.
Verde and Ivory Coast (Fig. 3 B, 3D). Both larval Our results shed light on the complex
retention (Fig. 3A) and dispersion are ecologically connective pathways among Tropical Atlantic spiny
relevant, the former contributing to the self- lobster stocks against which contemporary gene flow
recruitment or self-seeding processes (e.g., Sammarco and historical biogeography hypotheses may be tested.
and Andrews, 1988; Almany et al., 2007), and the The genetic structure and variability between close
latter to the connectivity among populations on larger and distant Tropical Atlantic pools of spiny lobster
scales than the ecological timescale. For example, stocks should also be addressed to investigate the
results of a density-independent matrix population influence of the simulated trajectories on the spatial
model suggest that a spatially structured fishery genetic heterogeneity (i.e., heterozygosity). Assuming
resource connectivity is as important as the that global climatic events such as the El Niño
productivity. Sink areas also may be essential for the Southern Oscillation do impact fisheries in many
persistence of a network (HASTINGS; BOTSFORD, different ways, it is important to preserve rare alleles
2006). to increase fitness under unusual environmental
The advection/diffusion model did not conditions (see Pérez-Ruzafa et al., 2006 and
demonstrate a clear connection between the eastern references therein).
and western Tropical Atlantic lobster populations.
However, the release of larvae from Ascension Island CONCLUSIONS
during El Niño simulations, resulted in the largest
zonal dispersal distance (Fig. 4B, 4D). In this case, The simulations performed show that
larval trajectories reach further east and southwest, extreme episodes such as those associated with strong
allowing the connection among populations of El Niño events could promote connectivity between
Ascension and those located in the northeastern Brazil adult spiny lobster stocks over the Tropical Atlantic,
section beyond the 5º S limit. Simulations also making possible a genetic flux by larval transport.
indicate connectivity of Ascension Island with the Results also indicate the occurrence of both retention
Fernando de Noronha Archipelago and the northeast and long-distance dispersal of spiny lobster larvae
of Brazil under the influence of the North Brazil within the Tropical Atlantic.
Current, northwards of 5º S. We suggest that these As a final remark, we conclude that the use
islands constitute a system of stepping-stones for of altimetry geostrophic current data, combined with a
marine species in the Tropical Atlantic with a long- relatively simple advection/diffusion model, can
lived larval phase, since they are positioned centrally provide the physical dynamics background required to
in the southern portion of the Tropical Atlantic. adequately address larval transport patterns. Synoptic
Looking at the predominantly unidirectional views of larval distribution for a range of
pattern of the trajectories of larvae released from the geographical, hydrodynamic and climatic conditions
Fernando de Noronha Archipelago (as well as from the on different spatial and temporal scales are useful to
Ivory Coast) one can appreciate the influence place fisheries management decisions within the
exercised by the geographical location of the spawning seascape/biogeographical context.
areas relative to both large scale, basinwide currents as
well as to mesoscale instability features. Adding to the
influence of mean advective fields, larval trajectories
ACKNOWLEDGMENTS
are also subject to seasonal changes and interannual
The authors would like to express their
variabilities such as those depicted by simulations
gratitude to Ramon Freitas, Dr. Jorge Conrado
starting in different periods of the year (April and
Conforte and Dr. Antônio Correia for their assistance
September) and during El Niño years.
with the model programming and Dr. Arcilan Assireu
The fact that we have extracted the weekly
for helping estimate the eddy-diffusion coefficients.
geostrophic fields for the El Niño period from the
This work was supported by CNPq-Brazil grants (no.
same database used for the 2000/2003 mean currents
131.758/04-7 and no. 384.462/06-5).
highlights the intrinsic variability captured by the
 16 BRAZILIAN JOURNAL OF OCEANOGRAPHY, 57(1), 2009

REFERENCES HASTINGS, A.; BOTSFORD, L. W. Persistence of spatial

populations depends on returning home. Proc. Natn.
Acad. Sci. U.S.A., v. 103, n. 15, p. 6067–6072, 2006.
ALMANY, G. R.; BERUMEN, M. L.; THORROLD, S. R.; HOLTHUIS, L. B. 1991. FAO species catalogue. Marine
PLANES, S.; JONES, G. P. Local replenishment of lobsters of the world: An annotated and illustrated
coral reef fish populations in a marine reserve. Science, catalogue of species of interest to fisheries known to
v. 316, n. 5825, p. 742-744, 2007. date. FAO Fisheries Synopsis, Rome, v. 125, n. 13, p.
ASSIREU, A. T. Estudo das características cinemáticas e 1-292, 1991.
dinâmicas das águas de superfície do Atlântico Sul LAGERLOEF, G. S. E.; MITCHUM, G. T.; LUKAS, R. B.;
Ocidental a partir de derivadores rastreados por NILLER, P. P. Tropical Pacific near surface currents
satélite. 2003. 174 p. Ph.D. Thesis. Instituto estimated from altimeter, wind and drifter data. J.
Oceanográfico da Universidade de São Paulo, São geophys. Res., v. 104 (C10), p. 23313-23326, 1999.
Paulo, 2003. Available for download at LEIS, J. M. Are larvae of demersal fishes plankton or
<http://www.ltid.inpe.br/dsr/grupos/hidrosfera/arquivos/ nekton? Adv. mar. Biol., v. 51, p. 57-141, 2006.
arcilan.pdf>. MONTGOMERY, J. C.; JEFFS, A.; SIMPSON, S. D.;
BECKER, B. J., LEVIN, L. A., FODRIE, F. J., MEEKAN, M.,; TINDLE, C. Sound as an orientation
MCMILLAN, P. A. Complex larval connectivity cue for the pelagic larvae of reef fishes and decapod
patterns among marine invertebrate populations. Proc. crustaceans. Adv. mar. Biol., v. 51, p. 143-196, 2006.
natn Acad. Sci. U.S.A., v. 104, n. 9, p. 3267–3272, PÉREZ-RUZAFA, A.; GONZÁLEZ-WANGÜEMERT, M.;
2007. LENFANT, P.; MARCOS, C.; GARCÍA-CHARTON, J.
BOOTH, J. D., PHILLIPS, B. F. Early life history of spiny A. Effects of fishing protection on the genetic structure
lobster. Crustaceana, v. 66, p. 271-294, 1994. of fish populations. Biol. Conserv. , v. 129, p. 244-255,
CAVALCANTE-SOARES, C. N.; FONTELES-FILHO, A. 2006.
A. 2000. Época de reprodução da lagosta-verde, PHILLIPS, B. F.; SASTRY, A. N. Larval Ecology. In:
Panulirus laevicauda (Latreille), no estado do Ceará, COBB, J. S.; PHILLIPS, B. F. (Ed.). The biology and
Brasil. [Spawning season of the smoothtail spiny lobster, management of lobsters, v. 2. New York: Academic
Panulirus laevicauda (Latreille), off Ceará, Brazil]. Arq. Press, 1980. p. 11-57.
Ciênc. Mar, v. 33, p. 43-50, 2000. PHILLIPS, B. F.; COBB, J. S.; GEORGE, R. W. General
CHISWELL, S. M.; BOOTH, J. D. Rock lobster Jasus Biology. In: COBB, J. S., PHILLIPS, B. F. (Ed.). The
edwardsii larval retention by the Wairarapa Eddy off biology and management of lobsters, v. 1. New York:
New Zealand. Mar. Ecol. Prog. Ser., v. 183, p. 227– Academic Press, 1980. p. 1-82.
240, 1999. PHILLIPS, B. F.; SASTRY, A. N. Larval Ecology. In:
CHISWELL, S. M.; WILKIN, J.; BOOTH, J. D.; COBB, J. S.; PHILLIPS, B. F. (Ed.). The biology and
STANTON, B. Trans-tasman Sea larval transport: Is management of lobsters, v. 2. New York: Academic
Australia a source for New Zealand rock lobster? Mar. Press, 1980. p. 11-57.
Ecol. Progr. Ser., v. 247,p. 173-182, 2003. PICAUT, J. Use of the Geostrophic approximation to
COELHO, P. A.; RAMOS-PORTO, M. Malacostraca - estimate time-varying zonal currents at the Equator. J.
Eucarida. Palinuridea. In: YOUNG, P. S. (Ed.). geophys. Res., v. 94(C3), p. 3228-3236, 1989.
Catalogue of Crustacea of Brazil. Rio de Janeiro: POLOVINA, J.; KLEIBER, P.; KOBAYASHI, D. R.
Museu Nacional, 1998, n 6. p. 387-392. . Application of TOPEX/POSEIDON satellite altimetry to
<http://acd.ufrj.br/mndi/Carcinologia/hp/Text/Palinurida simulate transport dynamics of larvae of spiny lobster,
e.htm>. Last access: Jan.16th, 2006. Panulirus marginatus, in the Northwestern Hawaiian
COWEN, R. K.; LWIZA, K. M. M; SPONAUGLE, S.; Islands, 1993–1996. Fish. Bull., v. 97, p. 132–143, 1999.
PARIS, C. B.; OLSON, D. B. Connectivity of marine SAMMARCO, P. W., ANDREWS, J. Localized dispersal
populations: Open or closed? Science, v. 287, p. 857- and recruitment in Great Barrier Reef Corals: the Helix
859, 2000. experiment. Science, v. 239, p. 1422-1424, 1988.
COWEN, R. K.; PARIS, C. B.; SRINIVASAN, A. Scaling STRAMMA, L.; RHEIN, M.; BRANDT, P.; DENGLER,
of connectivity in marine populations. Science, v. 311, p. M.; BÖNING, C. WALTER, M. Upper circulation in
522-527, 2006. the western Tropical Atlantic in boreal fall 2000. Deep-
FOLTZ, G. R.; GRODSKY, S. A.; CARTON , J. A.; Sea Res., I , v. 52, p. 221-240, 2005.
MCPHADEN, M .J. Seasonal mixed layer heat budget of SUGDEN, A., PENNISI, E. When to Go, Where to Stop.
the Tropical Atlantic Ocean. J. geophys. Res., v. 108, Science, v. 313 , n. 5788, p.775, 2006
15-1, 15-13, 2003. TAVARES, M. Lobsters. In: FISCHER, W. (Ed.). FAO
FREITAS, R.; CASTRO, M. Occurrence of Panulirus argus species identification sheets for fishery purposes:
(Latreille, 1804) (Decapoda, Palinuridae) in the Western Central Atlantic (Fishing Area 31), 1ed. Roma:
northwest islands of the Cape Verde Archipelago FAO, 2003. p. 292-325,
(Central-East Atlantic). Crustaceana v. 78, n. 10, p. ZENK, W., KLEIN, B., SCHRODER, M. Cape Verde frontal
1191-1201, 2005. zone. Deep-Sea Res. 38: 505-530, 1991.
GRIFFIN, D. A.; WILKIN, J. L.; CHUBB, C. F.; PEARCE,
A. F.; CAPUTI, N. Ocean currents and the larval phase
of Australian western rock lobster, Panulirus Cygnus.
Mar. Freshw. Res., v. 52:, p. 1187-1199, 2001.
GRIMM, V. Ten years of individual-based modeling in (Manuscript received 23 November 2007; revised
ecology: What have we learned and what could we learn 16 April 2008; accepted 11 September 2008)
in the future? Ecol. Model., v. 115, p. 129-148, 1999.