Journal of Marine Systems 54 (2005) 3 – 37
www.elsevier.com/locate/jmarsys
Coupling between the open ocean and the coastal upwelling
region off northwest Africa: water recirculation and
offshore pumping of organic matter
J.L. Pelegrı́*, J. Arı́stegui, L. Cana, M. González-Dávila, A. Hernández-Guerra,
S. Hernández-León, A. Marrero-Dı́az, M.F. Montero, P. Sangrà, M. Santana-Casiano
Facultad de Ciencias del Mar, Universidad de Las Palmas de Gran Canaria, 35017 Las Palmas de Gran Canaria, Spain
Received 24 May 2002; accepted 1 July 2004
Available online 29 September 2004
Abstract
The surface and upper-thermocline waters of the Canary Basin are characterised by very strong coupling between the open
ocean and the coastal upwelling region. Such coupling has its origin in water inflow into the upwelling region north of the
Canary Islands and its recirculation south along the continental slope, which is the true Canary Current. A portion of this
recirculating water is intermittently exported offshore through surface filaments. During late fall, a major diversion takes place
at Cape Ghir, allowing the presence of northward flow from Cape Blanc till Cape Yubi. The fraction of water that flows through
the Canary Archipelago is the origin of intense mesoscale variability south of the Canary Archipelago, which interacts strongly
with the coastal region. These physical characteristics are responsible of intense alongshore and vertical fluxes of nutrients and
dissolved inorganic carbon within the upwelling region. Coastal filaments and cyclonic eddies cause localised offshore export
of nutrients and organic matter, making possible that respiration be several times larger than production in the open ocean. A
major characteristic of the ecosystem comes from the seasonal variation in the current pattern, allowing coastal convergence and
intense transfer of coastal properties to the open ocean during late fall.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Coastal upwelling; Canary Current; Water recirculation; Organic matter
1. Introduction
* Corresponding author. Current address: Institut de Ciències
del Mar, CMIMA-CSIC, Grup d’Oceanografia Fisica, Passeig
Marı́tim de la Barceloneta, 37-49, Barcelona, 08003, Spain. Fax:
+34 932 309555.
E-mail address: [email protected] (J.L. Pelegrı́).
0924-7963/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmarsys.2004.07.003
Nutrient supply, primary production, organic matter export and nutrient mineralization constitute the
basic sequence for completion of the biogeochemical
cycle. The spatial variability of this cycle depends on
the physical mechanisms that operate at local and
4
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
regional scales. In some areas the physical environment is capable of supporting large nutrient input that
leads to high productivity. The best example of such
areas is the upwelling regions in the eastern boundary
of all subtropical gyres where the wind regime
accelerates the nutrients influx, allowing high rates
of primary production, and carbon recycling and
export (e.g. Huntsman and Barber, 1977; Codispoti
et al., 1982; Schulz, 1982; Jewell, 1994). In upwelling
regions, however, additional mechanisms may favour
the persistence of biogeochemical transformations,
even with weak or no upwelling favourable winds.
Localised near-horizontal water exchange between the
interior and coastal oceans for example, may bring
large amounts of nutrients into the coastal region and
provide paths for carbon export. And the specific
planktonic community structures may be adapted to
the changing physical environment and intermittent
nutrient input in the upwelling regime.
Although early ideas of the recirculation of
subtropical gyres proposed a weak diffuse eastern
flow, more recent circulation maps of the North
Atlantic Ocean illustrate an eastward flowing Azores
Current that splits into several rather energetic southward branches (Schmitz and McCartney, 1993; Reid,
1994). The number, intensity and location of most of
these branches seem to be highly variable, their origin
likely linked to developing instabilities of the Azores
Current. The easternmost branch, however, always
flows east into the coastal upwelling region off
northwest Africa and is what we identify as the
Canary Current. This branch is possibly the best
available example of large-scale interaction between
the coastal upwelling and open ocean regions.
Several regional studies, or studies of the Atlantic
circulation with enough data resolution in the eastern
boundary, have indeed improved our view of the
Canary Current (Stramma and Siedler, 1988; Lozier et
al., 1995; Siedler and Onken, 1996). These studies
have confirmed the existence of water inflow from the
open ocean into the coastal upwelling region north of
the Canary Islands. This may be interpreted as the
coastal upwelling region being the true eastern
boundary condition for the subtropical gyre, and the
subtropical gyre providing the continuous large-scale
forcing on the coastal ocean. How the system is
coupled however, is not easily distinguishable in these
early studies because they lacked enough spatial
resolution in the coastal upwelling region. The
importance of this coupling, emphasised by several
authors (Pelegrı́ et al., 1997; Barton, 1998; Barton et
al., 1998), lead to its study in a rather large European
Union project (CANIGO). In CANIGO several works
dealt with the coastal transition zone north of the
Canary Archipelago, where the water recirculates
south along the continental slope, as well as with
specific coastal sites where quasi-permanent filaments
stretch offshore (Hernández-Guerra et al., 2001, 2002;
Laiz et al., 2001; Pelegrı́ et al., 2005).
Our objective here is to examine the large-scale
and mesoscale processes that develop in the open
ocean and coastal upwelling coupled system, and to
scrutinise which mechanisms support the biogeochemical processes in the surface and upper thermocline layers of this region. Specifically, we want to
examine what the physical (wind forcing, solar
radiation, influx from the open ocean, recirculation
through the upwelling region, stirring role of the
Canary Archipelago) and biological (nutrients availability, plankton community structure, productivity)
mechanisms are that control the marine carbon system
in the upper thermocline waters off northwest Africa
(particulate and dissolved organic carbon, partial
pressure of CO2, total alkalinity, pH, total inorganic
carbon). These processes act at different scales, from
nutrient influx at the basin scale (Canary Current) to
increased productivity and carbon export in mesoscale
structures (filaments stretching out of the coastal
upwelling region and vortices formed at the lee side
of the Canary archipelago). One important point is
that local processes are the response to both local and
global forcing, the character of the biogechemical
cycle eventually relying on the energy found within
the ocean interior and the rapid exchange of properties
in the boundary region.
While there have been a number of recent reviews
dealing with several specific characteristics in the
eastern boundary of the subtropical gyre, our
approach is clearly interdisciplinary. We are interested
in interconnecting those processes that develop in the
surface and upper-thermocline layers, responsible for
sustaining the biogeochemical cycle. For this reason,
we have not attempted to be exhaustive on each
specific component of this complex system and, at
each section, the reader will be directed to the relevant
references. The geographic region to be examined,
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
both in the horizontal and vertical directions, has also
been rather constrained as compared with other work
in the literature. Our region of interest is the surface
and upper thermocline waters of the Canary Basin,
including their possible association with the poleward
undercurrent and the Antarctic Intermediate Water at
the eastern boundary. This spans a vertical extension
from the sea surface down to approximately 750 m,
and a horizontal area from the African Coast to about
208W and from the Strait of Gibraltar to the lee side of
the Canary Islands. These meridional limits include
the Canary Current, cruising into and along the
African continental slope, and the complex pattern
of mesoscale structures that arise south of the Canary
Islands because of the interaction of the Canary
Current with this archipelago.
2. Air–sea interaction
2.1. Atmospheric forcing
The zonal available potential energy associated
with latitudinal temperature gradients is the main
energy source in middle latitude systems. In the
tropics, on the other hand, the storage of available
potential energy is comparably small due to the much
weaker temperature gradients, making ocean–atmosphere heat exchange the primary energy source for
atmospheric disturbances, especially for those that
originate within or near the equatorial zone (e.g.
Palmén and Newton, 1969). Here we briefly review
the annual cycle of air–sea heat flux in the Canary
Basin as well as the dominant wind patterns and their
temporal and spatial variability. One aspect to keep in
mind is that the circulation patterns are the response to
both mean and intermittent wind forcing. Localised
features, such as coastal upwelling, have a rather fast
response, so we may expect that these will be able to
react to the transient forcing mechanisms. Large-scale
features, such as the intensity and location of the
Canary Current, have a much greater inertia and
indeed will depend on the combined effect of largescale mean and intermittent forcing.
The temporal variation of short-wave radiation
(SW), net long-wave radiation (IR), latent heat flux
(LE) and sensible heat flux (H) may be used to
characterise the rate of air–sea heat exchange around
5
the Canary Archipelago. Fig. 1 shows the temporal
variation of these energy fluxes in the Canary
Archipelago region as obtained from the climate atlas
by Isemer and Hase (1987). Fig. 1b shows the annual
cycle of the net radiation flux (Net), the oceanic heat
loss (Hl) and the net air–sea heat flux (Hf) in W m 2.
The net radiation flux is the difference between
incoming SW and outgoing IR; the oceanic heat loss
is calculated from the sum of IR, LE and H; and the
net air–sea heat flux is calculated using the data of
SW, IR, LE and H. In the Canary Archipelago region,
the net heat flux is positive from March to August and
negative from September to February.
To calculate SW, we follow Reed’s (1977)
formulae, applicable for daily averages. The expression uses incident solar radiation at the surface,
cloud cover, noon solar altitude and clear sky
radiation to evaluate the short-wave radiation.
Minima are closely related to maxima of total cloud
cover and vice versa (290 W m 2 in June, 110 W
m 2 in December). The second major component
determining the net air–sea heat flux is LE. This flux
is usually directed upwards, providing for main
ocean loss. The patterns of LE are determined by
the wind speed and mixing ratio, presenting a
relative maximum of 220 W m 2 in the trade wind
zone. The most constant among all air–sea heat flux
components is IR, its magnitude being mainly
influenced by total cloud cover, air temperature
and humidity, with the air–sea surface temperature
difference playing a minor role. In the upwelling
region off northwest Africa, the maximum net longwave radiation (50 W m 2) occurs in January,
associated to a minimum of cloudiness. Sea surface
temperature (SST) and surface wind speed dictate
the seasonal pattern of H. This flux is relatively
small during the complete annual cycle (30 W m 2
in December, 5 W m 2 in June), particularly if
compared with its value in the western and northwestern part of the Atlantic Ocean.
Schmitt et al. (1989) compiled evaporation and
precipitation data for the whole North Atlantic Ocean
to calculate the haline density flux into this ocean. In
the Canary Basin, mean annual evaporation exceeds
precipitation by values ranging between 50 cm year 1
in its northern portion and over 100 cm year 1 in the
southern portion. One important characteristic is that
the difference remains almost constant throughout all
6
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Fig. 1. Seasonal variation of heat flux in the Canary Island region (after Isemer and Hase, 1987). (a) Net short-wave radiaton (SW), net long-wave
radiation (IR), sensible heat flux (H), and latent heat flux (LE). (b) Net radiation (Net), oceanic heat loss (Hl), and net air–sea heat flux (Hf).
year, in contrast to what happens over most of the
North Atlantic Ocean. Schmitt et al. (1989) further
compared the haline and thermal density fluxes and
obtained that the thermal density flux exceeds the
haline flux by a factor between 2 and 20 in the Canary
Basin, with these values increasing towards the
African coast. The relatively high evaporation rate
in the open ocean is related to the rather large and
consistent net air–sea heat fluxes and high sea-surface
temperature values. Minimum values occur in May
(1.3 cm month 1 in Gran Canaria island) and
maximum values take place in December and January
(1.9 cm month 1 in Gran Canaria island). Total
rainwater measured at this same location is 16.5 cm
year 1, confirming that the Canary region is a
geographical area with strong evaporation.
The wind regime in the Canary Basin is mainly
determined by the location of the Azores high, which
causes that the mean wind direction is from the
northeast all year long with changes being related to
synoptic and mesoscale weather systems (INM, 1988).
The most characteristic feature of the summer pattern
is the ubiquitous Azores high-pressure elliptical centre
that, together with the north Africa low, produces the
well-defined northeasterly trade wind regime. In the
Canary Archipelago the trade winds reach maximum
monthly mean values of 8.7 m s 1 in August. During
autumn or spring, the Azores anticyclone is not as
deep as in summer, typically presenting a quite
elongated shape. The winter regime is determined by
the southern movement of the Azores high, which
causes the passage of low-pressure centres through the
Canary Basin to be a common feature. During this
season, however, mean winds are still from the
northeast, the minimum monthly mean values corresponding to January (2.7 m s 1).
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Plate 1. SST satellite images for (a) August 27, 1998, (b) November 19, 1998, and (c) March 20, 1998. The colour temperature code is shown in a vertical scale at the right of the
images (units in 8C). The insets show the corresponding surface pressure fields (units in mbar).
7
8
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Fig. 2. (a) Temporal evolution of the top 200 m at stations located 100 km north of Madeira (top) and 20 km east of Lanzarote (bottom), with the
thick line showing the annual average and the thin line the actual temperature profile (the location of both stations is shown in Fig. 4). (b)
Idealised sketch illustrating the seasonal evolution of the surface mixed layer, with the one-/two-letter code indicating the season, for example
late-spring (ls), reproduced from Ratsimandresy et al. (2001).
Plate 1 presents SST images for 3 days, each
corresponding to a different season, that show the
latitudinal shift of isotherms in the surface layers as a
response of heat accumulation or loss. Insets in this
plate show the concurrent surface pressure fields,
which illustrate the characteristic local patterns of
surface pressure for summer (Plate 1a), autumn or
spring (Plate 1b), and winter (Plate 1c). The pressure
patterns are consistent with the ubiquitous presence of
northeasterly winds from spring into fall (Plate 1a and
b), with a maximum intensity during summer. These
winds reinforce the Canary Current throughout these
months, which attains its maximum intensity by late
autumn. During winter the mean wind conditions
respond to the high pressure centre located in its
southernmost position, still causing northeasterlies but
much weaker than during summer (Plate 1c). At this
time, it is frequent to find southwesterlies, associated
to the passage of low-pressure centres, which cause
important forcing fluctuations in relatively short time
scales, a few days as compared with a characteristic
time scale of 10 days during summer.
Fig. 3. Seasonal evolution of (a) temperature and (b) temperature anomaly in the uppermost 200 m in a transect Gran Canaria–Madeira–Lisbon
(adapted from Ratsimandresy et al., 2001). See Fig. 4 for the location of this transect.
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
9
10
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
2.2. Surface mixed layer
Below 200 m, the temperature distribution experiments only minor seasonal variations. The major
changes in temperature take place in the topmost 200
m of the water column, following heat exchange
between the ocean and atmosphere and the deepening of the surface mixed layer, the latter mainly
resulting from increased summer winds and winter
vertical convection. A good description of the
seasonal evolution of the temperature in the uppermost layers of the water column has been provided
by several authors (Barton et al., 1998; Ratsimandresy et al., 2001). Fig. 2, reproduced from Ratsimandresy et al. (2001), illustrates an idealised sketch
of the temporal evolution of the mixed layer. The
mixed layer is deep and cold in March, following
winter ocean heat loss and deep convection. In June
surface heating causes that the surface waters are
capped with a shallow warm layer, which increases
its thickness throughout summer due to the action of
the intense winds. After October the ocean starts
losing heat and the surface temperature decreases
while the surface layer thickness continues to
increase.
Fig. 3, also reproduced from Ratsimandresy et al.
(2001), illustrates the seasonal evolution of the
temperature and temperature anomaly in the uppermost 200 m of a transect from the Canary Islands to
Madeira. The seasonal evolution of the surface mixed
layer is clearly visible from the temperature plots (Fig.
3a). The temperature anomaly plots, on the other
hand, illustrate the relative warming and cooling of
the surface layers as compared with the annual mean
situation (Fig. 3b). A remarkable feature in the
temperature anomaly plots is the subsurface (between
50 and 100 m) maxima easily visible during winter.
Siedler et al. (1987) suggested that this water, left
below the seasonal thermocline, is the origin of about
1 Sv of Madeira Mode Water that finds its way to
greater depths while shifting westward. This amount
of modal water, however, is still rather small as
compared with the Eighteen-Degree water formed in
the western subtropical gyre (Worthington, 1959).
One important aspect in the Canary Basin is that
the annual mean surface winds cause water convergence at the sea surface (wind-influence) layer, which
leads to negative Ekman pumping and water sub-
duction into the ocean interior. Stommel (1979)
indicated that, because of the latitudinal motion of
the surface layer, surface waters only escape to the
ocean interior during winter. Hence, the amount of
water that does escape from the surface depends on
both winter mean conditions and the passage of
atmospheric frontal systems. As discussed above,
despite its seasonal heat gain–loss cycle, the region
has a net heat gain. The main reason for such gain is
that the Canary Current is a relatively cool current,
particularly cold in the upwelling region, susceptible
to gain heat. The rate of heat gain in the Canary Basin
has been estimated by Pelegrı́ et al. (1997) to be of
order 0.05 PW. The escape of some of this heat from
the ocean surface, either through the formation of
modal water or by subduction of relatively salty water,
could be an efficient mechanism for the ocean interior
to absorb heat excess generated in the atmosphere
(Bindoff and McDougall, 1994).
3. The open ocean
3.1. Water masses
In the frame of the CANIGO project, a closed box
was carried out four times (January 97, September 97,
April 98 and July 98) spanning the region between the
African coast, the Canary Islands (298N) and a
parallel near Madeira (328N). The objective was to
determine the mass, heat and freshwater balances in
this region. Some of these sections have been recently
analyzed and the results are progressively being
published (Pérez et al., 2001; Knoll et al., 2002).
The southern section (the 298N section in Fig. 4) is
used here to describe the principal water masses
present in the area. The vertical distribution of the
mean isohalines, obtained by averaging all four
cruises, is most helpful for this purpose because of
the very different salinity signature of the water
masses (Fig. 5).
In the surface layer, from 0 to approximately 700m depth, we find North Atlantic Central Water
(NACW). This water mass is characterised by nearly
horizontal isohalines (Tomczak, 1981), as observed in
Fig. 5. Several authors have differentiated NACW of
subtropical and subpolar origin, the latter with salinity
in this region over 35.66 (Harvey, 1982; Pollard and
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
11
Fig. 4. The Canary Basin, with major bathymetric and geographical features. The solid and dotted lines refer to sections discussed in the text,
while the crosses and solid boxes refer to stations also discussed in the text. The dashed horizontal areas correspond to upwelling regions and the
divot area corresponds to the region where systems of cyclonic and anticyclonic filaments are present. The arrows represent the major surface
filaments that emanate offshore from the upwelling region.
Pu, 1985; Pérez et al., 2001). From Fig. 5 we
appreciate that NACW of subtropical origin occupies
the top 500 m while NACW of subpolar origin
occupies the depth range between 500 and 750 m. Our
main interest here is the surface mixed layer and
subsurface waters down to 750 m, and we will refer to
these subsurface waters indistinctly as the upperthermocline or NACW.
At intermediate layers (from approximately 600- to
1500-m depth), two water masses are present:
Antarctic Intermediate Water (AAIW), detected by
its low salinity (b35.3), and Mediterranean Water
(MW), with high salinity values (35.5) (HernándezGuerra et al., 2001). AAIW is clearly observed in the
eastern boundary (up to 600 m) as a rather spread
tongue, much wider than the classical alongshore
poleward current (Barton, 1998; Pérez et al., 2001).
Since Fig. 5 has been obtained using the mean of all
four cruises, the existence of these minimum salinity
values confirms that AAIW is a permanent feature at
12
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Fig. 5. Salinity distribution in a zonal section at 298N (see Fig. 4 for the location of this section).
the boundary. The MW is observed in the remaining
of the section. Several patches of water with
intermediate salinity are presumably due to mixing
of AAIW and MW. Under the MW (approximately
below 1500 m), we find North Atlantic Deep Water
(NADW), with salinity decreasing with depth and the
isohalines having a quasihorizontal distribution.
3.2. Chemical properties
Recent measurements of the marine carbon system
variables have been carried out in the Canary Region
in response to an increased interest in global carbon
change and greenhouse warming. The capacity of the
oceans to uptake carbon dioxide (CO2) depends
greatly on many factors such as hydrography, water
circulation, mixed layer dynamics, wind stress and
biological processes in the oceans (Broecker and
Peng, 1982). A systematic study of the dissolved
inorganic carbon system in the Canary Basin started in
October 1995, one year after setting up the ESTOC
time series station north of Gran Canaria (29810VN,
15830VW). At this location the partial pressure of
carbon dioxide (expressed as fugacity, fCO2) shows a
seasonal pattern in the surface waters characterised by
CO2 dissolution from December to May and outgassing from June to November (González-Dávila and
Santana-Casiano, 1999; Santana-Casiano et al., 2001).
Considering the CO2 gas transfer coefficient of
Wanninkhof (1992) and a mean wind of 7 m s 1, it
turns out that on an annual scale the region acts a
small sink of CO2, at a rate of 0.15 mol m 2 year 1
(Santana-Casiano et al., 2001; González-Dávila et al.,
2003). This value is found in the range estimated by
Takahashi et al. (1999), between 0 and 1 mol m 2
year 1.
The distribution of carbon-related variables in the
Canary Basin presents several characteristic features
associated with the different water masses. AAIW is
found close to the African shelf-break, characterised
by the lowest pHT values (total hydrogen pH scale,
normalized at 25 8C) in the Canary region, around
7.62. The signal of this water is traced by its pH along
a zonal transect from 208W to the African Coast (very
close to the 298N section in Fig. 4), and along the
coast up to a latitude beyond 308N, in front of Cape
Ghir (Santana-Casiano et al., 2001). The northward
transport of AAIW has also been identified by Pérez
et al. (2001) from the analysis of nutrients, dissolved
oxygen, alkalinity and dissolved inorganic carbon.
The influence of the Mediterranean outflow Water
(MW) is observed north of the Canary Archipelago at
depths about 1100–1200 m, identified by a high value
of total alkalinity (around 2360 Amol kg 1) (Pérez et
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
13
representative values for the chemical characteristics
of some of these water types. In the upper-thermocline, down to about 750 m, there is only NACW of
subtropical and subpolar origin (although Pérez et al.,
2001 introduce a third water type, which they named
H, identified as the limit between these two water
types). An exception is the eastern boundary where
AAIW may be found below about 500 m, in a
maximum percentage of 11%. For this reason, Pérez
et al. (2001) introduced a new water type (AA) that
contained the influence of NACW of subpolar origin
and AAIW, characterised by large values of nutrients.
They suggest that this water is rapidly diluted further
north where any observed large nutrient concentration
is caused by regeneration of organic matter associated
to upwelling processes.
An important hypothesis posed by Pérez et al.
(2001), on the basis of historical and recent data sets,
is that during the last decades there has been a
progressive decrease in the northward penetration of
AAIW. It is possible, however, that the analysis of the
scarce available data does not reflect a decadal
tendency but simply a seasonal variability. This
alternative explanation is based on the relation
between the poleward undercurrent and the presence
of AAIW and the poleward undercurrent. The
hypothesis, to be discussed below, is that there is a
seasonal cycle in the poleward undercurrent, responding to seasonal changes in the preferential southward
path of the Canary Current.
al., 2001). MW water has also been observed north of
La Palma Island in rotating lenses of high salinity and
alkalinity (Santana-Casiano et al., 2001).
Nutrients are almost undetectable all year long
within the mixed layer, except during winter when
deep convective mixing incorporates nutrients from
the upper-thermocline. At this time, nitrate concentration may increase to 1–2 Amol kg 1 in the top 150
m (Pérez et al., 2001; Pérez-Rodrı́guez et al., 2001).
Phosphate concentration follows this same pattern
while the silicate distribution in surface water shows
intermittent high values (about 2 Amol kg 1) due to
Sahara Dust events arriving to the region. Below the
mixed layer we find the nutricline, where nutrient
concentrations roughly increase with potential density
except in those regions where the isopycnals approach
the surface and nutrients are utilised. This correlation
is indicative of southwestward water recirculation
along isopycnals, following water subduction in
regions where Ekman pumping is negative (Leetma
and Bunker, 1978; McClain and Firestone, 1993). As
the upper-thermocline waters recirculate, dissolved
nutrients increase through the dissolution of falling
particles and regeneration of organic matter (Kawase
and Sarmiento, 1985). Nitrate is very well linearly
correlated with phosphate throughout the whole water
column. Nitrate is nonlinearly correlated with silicate
within NACW due to redissolution of opal skeletons
and decreased regeneration of nutrients with depth,
particularly near the upwelling waters at the eastern
boundary (Pérez et al., 2001).
Pérez et al. (2001; see also Castro et al., 1998) have
described the chemical and nutrient characteristics of
the eastern North Atlantic in terms of water masses,
with their relative proportions obtained using a mixing
model of temperature, salinity, nutrient, oxygen and
carbon system variables. Table 1 shows a summary of
3.3. The seasonal plankton production cycle
In the open ocean waters of the Canary region, the
annual plankton production cycle is affected by the
quasi-permanent seasonal thermocline. This sharp
density transition restricts the vertical flux of inor-
Table 1
Characteristic values of chemical properties for water types in the upper-thermocline of the Canary Basin (adapted from Pérez et al., 2001)
Type/property
S
T
O2
CT
AT
pHT
NO3
SiO2
PO4
ENACWt
H
ENACWp
AA
36.672
35.660
35.230
34.900
18.50
12.20
8.56
6.50
229F1
194F1
157F2
119F4
2090F1
2148F1
2189F2
2218F3
2397F1
2336F1
2321F1
2312F1
8.014F0.002
7.818F0.002
7.695F0.003
7.592F0.005
0.4F0.2
13.7F0.2
24.5F0.3
29.7F0.6
0.3F0.2
4.9F0.2
14.5F0.4
19.9F0.7
0.01F0.01
0.79F0.01
1.55F0.02
1.92F0.04
ENACWt (Eastern North Atlantic Central Water of subtropical origin); ENACWp (Eastern North Atlantic Central Water of subpolar origin); H
separates ENACWt from ENACWp; AA (Antarctic Intermediate Water end member). CT stands for total inorganic carbon, AT for alkalinity and
pHT for pH normalized at 25 8C.
14
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
ganic nutrients from upper thermocline waters to the
euphotic zone, limiting phytoplankton growth during
most of the year. As indicated above, the seasonal
thermocline only disappears during winter as the
result of surface cooling and the mixed layer reaches
its maximum penetration depth (150–200 m, see Fig.
3). The seasonal thermocline begins to reform in
April–May, leading to the more common situation of a
surface euphotic zone depleted of nutrients. Therefore,
it is only during the short mixing period that follows
deep convection when phytoplankton will grow fast.
In this circumstance, the phytoplankton builds up a
bbloom,Q the only condition being that cellular growth
rate be higher than the grazing rate of zooplankton.
The dynamics of this blate-winter bloomQ is more
similar to the autumn bloom (nutrient-limited) than to
the spring bloom (light-limited) in temperate seas.
From late spring to early winter, most chlorophyll
is found in a deep maximum (DCM) within the
seasonal thermocline, and large phytoplankton cells
are almost absent from surface waters. Indeed, small
picophytoplankton (b2 Am)—characteristic of oligotrophic regions—is the most abundant autotrophic
group, contributing to total chlorophyll and production more than 75% in autumn and up to 50% in
spring (Montero, 1993; Arı́stegui et al., 2001). These
cells, together with the heterotrophic bacteria and their
consumers, the heterotrophic flagellates, are the main
components of the autotrophic and heterotrophic
biomasses, respectively, of the planktonic food webs.
They are also mainly responsible for the remineralization of organic matter in surface waters (Arı́stegui
and Montero, 2005). Nevertheless, the structure of
this microbial community displays seasonal and
spatial changes, to a large extent related with
mesoscale variability induced by the flow passing
through the islands and the nearby coastal upwelling.
Only during late winter peaks in chlorophyll and
primary production may be observed in surface
waters. These maxima (0.5–1.0 mg m 3) are
significantly higher than the annual means (b0.2
mg m m 3; Braun and Real, 1984), but much lower
than the maxima (N2 mg m 3) observed in temperate waters during the spring bloom. Whether these
values are the result of the injection of nutrients into
the impoverished euphotic zone (resource control),
the effect of the grazing pressure (consumer control), or both, is still a matter of debate. Arı́stegui et
al. (2001) studied the short-term (days to weeks)
temporal variation in a coastal planktonic community at Gran Canaria Island, to asses the effect of
resource and consumer controls on the variability of
the seasonal planktonic cycle. They observed a clear
decoupling between the biomass and production of
autotrophs as well as an inverse relationship
between the biomasses of phytoplankton and mesozooplankton during the early stages of the bloom
session. This led them to suggest that primary
production could depend on a resource control
(nutrients) while biomass could depend on a
consumers (grazing) control. In oceanic waters the
year round variation in chlorophyll concentrations,
integrated over the whole water column, ranges
between 15 and 60 mg Chl m 2. This range of
variation is relatively small as compared with the
much larger variability in primary production values,
at least one order of magnitude greater (Basterretxea, 1994). This supports the hypothesis that phytoplankton biomass is mainly controlled by grazing
pressure.
Zooplankton biomass values are rather low
during most of the annual cycle, although they
depend on the proximity of the sampling site to the
African coast and whether it is on the windward or
leeward portion of the islands. North of Tenerife
Island, the average biomass value between 0- and
200-m depth was 390 mg dry weight (dw)d m 2
(Braun, 1981), while south of Gran Canaria Island
the average value for the annual cycle was 1370 mg
dwd m 2 for the same depth strata (Hernández-León
et al., 1984). This sharp difference is related to an
island mass effect on the biomass values of
plankton. A mesozooplankton bloom appears to
coincide with increased mixing during late winter
(Hernández-León, 1998; Hernández-León et al.,
1984, 1998, 2001c), although an increase of
biomass may be observed in July–August coinciding
with the maximum intensity of the Trade Winds on
the southern shelf of Gran Canaria Island (Hernández-León, 1988).
Mesozooplankton grazing in the Canary Island
waters is relatively small, in agreement with
observations for other oceanic areas. Based on
indices of metabolism and assuming herviborous
feeding, Hernández-León et al. (1998) estimated that
this potential grazing accounted for 10% to 46% of
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
the primary production. Arı́stegui et al. (2001)
found that mesozooplankton controlled 20% of
primary production in a coastal station, while
Hernández-León et al. (1999) found a value of
46% in the tropical oceanic waters. However, direct
estimations of grazing using the gut fluorescence
method showed values ranging from 11% to 22% of
primary production (Hernández-León et al., 2001c)
while ingestion (obtained from metabolism and
growth indices) accounted by 20–37%. Up to 41%
of the ingestion by mesozooplankton was nonpigmented food. Therefore, most of the grazing
control in these waters should be due to protista.
This is in agreement with the calculations made by
Dam et al. (1995) for the tropical Pacific where
most of the grazing is exerted by microzooplankton.
Finally, the control of zooplankton and protista on
primary production can change depending on the
bottom-up and top-down control in the rather complex
food web of warm water ecosystems. Recently,
Hernández-León (1998) and Hernández-León et al.
(2001b) found a lunar cycle in zooplankton abundance similar to the one described for freshwater
ecosystems (Gliwicks, 1986). A decrease in the
abundance of copepodites plus adult copepods was
observed from the second to the fourth quarter of the
moon. During the full moon the large zooplankton
and micronekton of the Deep Scattering layers (DSL)
do not reach the upper mixed layer, in order to avoid
predation under a relatively high level of illumination
(Pinot and Jansá, 2001), and epipelagic zooplankton
abundance increases due to the relatively low predatory pressure. On the other hand, during the new
moon phase the diel vertical migrants reach the
surface waters and prey upon epizooplankton, producing and abundance decrease. This lunar cycle in
zooplankton could explain the 30-day period observed
in sediment trap data in subtropical waters (Khripounoff et al., 1998). The effect of the diel vertical
migrants promotes not only the intensity of the active
flux in the ocean (see Hernández-León et al., 2001c
for the assessment of this flux around the Canary
Islands) but could also drive the variability of the
gravitational flux. Moreover, zooplankton is the food
of most larval stages of marine organisms and the
described variability can also explain the lunar cycle
usually observed in fish larvae and invertebrates in
warm water ecosystems.
15
4. Coastal upwelling recirculation
4.1. Along and cross-shore circulation
Recent observations have led to the viewing of the
Canary Current as a quite intense and well-localised
boundary current, with its behaviour being intimately
connected with the coastal upwelling region. The
coastal upwelling region is maintained by the presence of northeasterly winds all year long. From Cape
Blanc to the Strait of Gibraltar (roughly from 20 to
358N) the winds are upwelling favourable all year
long, although north of the Canary Archipelago they
are intense during the Summer months and moderate
or weak during all other seasons (Wooster et al., 1976;
Speth and Detlefsen, 1982; Nykjaer and Van Camp,
1994). From Cape Ghir to Cape Yubi (roughly 28 to
318N), the coastline presents a prominent concave
shape. This, together with the morphology of the
Atlas range in this region, causes the upwelling to be
much reduced, even during summer. This breakage in
coastal morphology and upwelling favourable winds
has, as we will see below, some potential important
consequences on the path of the Canary Current.
Upwelling is characterised by the presence of two
main dynamic features: a vertical cell and alongshore
currents (e.g., Allen, 1973; Brink, 1983). The vertical
cell is induced by the alongshore winds; these produce
offshore Ekman transport of surface waters that are
replaced by relatively cold subsurface waters, approximately flowing along the isopycnals. If strong
enough alongshore winds are maintained for a few
days, then surface waters are retrieved from the
coastal region and a frontal signature, marking a
contrast between the surface layer and subsurface
waters, shows off away from the coast (Csanady,
1977). The front is less visible in winter, because of
the absence of the seasonal thermocline. The upwelling system responds rapidly to the upwelling favourable winds. In times of about only 1 day a baroclinic
jet, in approximate geostrophic balance, sets up in the
offshore side of the upwelling front. This jet reaches
velocities up to 0.5 m s 1, but these are limited to the
upper layers ( ~50 m), so the total transport over a
width of the order of the internal Rossby radius of
deformation (~20 km) is typically less than 0.5 Sv
(Allen, 1973; Brink, 1983; Pelegrı́ and Richman,
1993). The baroclinic jet produces mixing between
16
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
the subsurface and surface layers, this vertical
exchange for water and momentum setting up an
offshore limit for the surface front. If the winds are
maintained for times of the order of 1 or several
weeks, then a relatively slow barotropic jet will
develop in the poorly stratified upwelled waters found
onshore of the surface front. This jet has large inertia
and will remain even if the upwelling conditions
temporally disappear. Furthermore, this barotropic
flow may sustain several times the water transport
associated to the baroclinic jet (Pelegrı́ and Richman,
1993).
This idealised two-dimensional picture is broken
because of the finite longitude of the coast and the
presence of irregularities in coastal morphology. The
presence of the Gibraltar Strait interrupts the meridional extension of the upwelling system, so the
alongshore baroclinic and barotropic transports
require a supply of open ocean water. In the
description of the eastern Atlantic circulation, this
corresponds to a sink of surface and upper-thermocline water that recurrently appears in different
analysis of historical data sets (Stramma and Siedler,
1988; Reid, 1994; Lozier et al., 1995; Siedler and
Onken, 1996). An example is shown in Fig. 6,
reproduced from Lozier et al. (1995), that shows the
mean streamlines at the 26.5 and 27.0 isopycnals,
typically found at depths of about 150 (just below
the mixed layer) and 400 m in the Canary Basin.
The speeds that correspond to these stream functions are about 1.5 and 0.7 cm s 1 into the coastal
region, and suggest a transport of several Sv into
the coastal upwelling region north of Cape Ghir.
This is consistent with some 3 Sv inferred by
Stramma (1984) for the top 800 m, and about 1 Sv
inferred by Stramma and Siedler (1988) for the top
200 m. The connection between the upper thermocline of the open ocean and the coastal upwelling
region is also evident in the distribution of potential
vorticity at the 26.5 and 27.0 isopycnals (Fig. 6c
and d, obtained from data provided courtesy of
Susan Lozier, Duke University; see also Lozier,
1997). The distribution in the deeper isopycnal
suggests a homogenization of potential vorticity in
the coastal upwelling region north of the Canary
Islands, hence possibly allowing the existence of
southward recirculation along the slope. These
results, however, were mostly obtained from open
ocean data so they are still unable to provide the
precise location and distribution of the water inflow.
More recently, Pelegrı́ et al. (unpublished manuscript; see Laiz et al., 2001, and Ratsimandresy et
al., 2001) have reported a mean flow of about 2 Sv
into the coastal upwelling region between Cape Ghir
and the Strait of Gibraltar.
The southward extension of the upwelling system
is limited by both the existence of an inflexion in
the African coastline, where winds are locally
modified, and the presence of the Canary Archipelago. These two factors are probably responsible for
the existence of the Ghir filament, a quasi-permanent surface structure that stretches offshore over
100 km near Cape Ghir (Hagen et al., 1996; Pelegrı́
et al., 2005). Pelegrı́ et al. (2005) estimate its net
offshore transport to be about 0.5 and 1 Sv at times
when the filament was weak and moderate, respectively. This appears to be only a fraction of the total
inflow north of Cape Ghir so the conclusion is that
most of the transport flows south, probably through
the eastern Canary Islands or between the Canary
Islands and the African coastline. The flow of the
Canary Current through the eastern Canary Islands
and the African coast has been estimated from
several years of XBT and moored current meter data
(Hernández-Guerra et al., 2002; Knoll et al., 2002),
as well as from historical tide gauge and hydrographic data (Navarro-Pérez and Barton, 2001). The
transport of the Canary Current flowing between
Gran Canaria Island and the African coast is
1.8F1.4 Sv southward. Approximately half of this
transport flows through the channel between the
islands of Gran Canaria and Fuerteventura and the
other half through the channel between Fuerteventura and the African coast. The transport presents a
significant seasonal variability, ranging between
1.2F0.3 Sv in May and 2.6F0.1 Sv in January.
The transport through the channels presents reversals (northward flow) during May (between Gran
Canaria and Fuerteventura) and November (between
Fuerteventura and Africa).
The major picture arising from the above
description is that the Canary Current flows south
along the upwelling transition zone past the Canary
Islands which constitutes the real eastern boundary
condition for the open ocean. Fig. 7 shows the
trajectories of surface buoys, dragged at a depth of
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
17
Fig. 6. Mean stream function for the (a) 26.5 and (b) 27.0 isopycnals (reproduced from Lozier et al., 1995). Solid contour intervals in the Canary
Basin are drawn every 0.5 m2 s 2. Thick lines indicate the locations where these isopycnals surface during winter and summer. (c) Logarithmic
distribution of the potential vorticity for the 26.5 isopycnal. (d) Logarithmic distribution of the potential vorticity distribution for the 27.0
isopycnal. The numbers provide the exponents of base 10 in units of 10 12 s 1 bar, so a given value p would correspond to 10 p 10 12 s 1 bar.
Data for c and d have been provided courtesy of Susan Lozier.
10 m, that were deployed at the Gulf of Cádiz on
February 1988. It is remarkable that these buoys did
not enter the Mediterranean Sea and followed a
swift southwestward trajectory, with speeds typically
over 0.1 m s 1, along the upwelling zone. This
trajectory was not modified by the offshore Ekman
transport associated to the northeasterlies at least
until past the Canary Archipelago. South of the
18
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Fig. 7. Trajectories of surface buoys illustrating the coastal jet (original data courtesy of Pierro Zanasca, Saclantc).
Canary Archipelago the buoys slowly drifted offshore, probably related to the great intensity of the
Ekman vertical cell. The trajectories in this southern
region are similar to trajectories of buoys deployed
at Cape Yubi in August 1999, which followed south
slowly drifting offshore, although when trapped by
coastal filaments the offshore movement became
very fast (Des Barton, personal communication).
A conceptual difficulty could arise when trying
to connect this quite permanent picture with the
intermittent character of upwelling. The permanence
of the southward flow probably lies on the mixed
baroclinic and barotropic character of the upwelling
jet, with the barotropic component being capable of
sustaining the southward transport even when
upwelling weakens. The intermittent character of
the upwelling zone is vividly illustrated by Plate 1,
which presents several satellite images of the SST
and the concurrent surface pressure fields. Plate 1a
illustrates how the upwelling signature (coastal cold
waters) is broken between Cape Yubi and Cape
Ghir, despite surface pressure maps suggest strong
upwelling favourable winds throughout the region.
In Plate 1b the upwelling signal is very well defined
from Cape Ghir until the west coast of the Iberian
Peninsula. In the spring image (Plate 1c), despite
the relatively weak winds, the SST signal provides a
sense of continuity along the coast.
The observations of winter flow reversals in the
upwelling zone near the Canary Archipelago are
supported by a study of seasonal variability by
Navarro-Pérez and Barton (2001). These authors
have used hydrographic and tidal gauge data to
describe a winter offshore diversion of the southward Canary Current as it approaches the Canary
Archipelago. Seasonal changes in regional circulation are based on the repartition of the Canary
Current in baroclinic and barotropic contributions.
The baroclinic flow responds to the seasonal pattern
of northeasterly winds, with maximum values taking
place during summer. At this time this flow may
often become unstable and depart from the coast as
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
surface filaments. On the other hand, the barotropic
component slowly builds up, from spring till fall,
with its maximum rate of change occurring during
the strong summer winds. By late fall, southward
transport is maximum and the flow becomes
unstable and separates from the coast, retaking its
southward motion further west between the central
or western Canary Islands. At this time the eastern
channels sustain northward flow, having its origin
on a large cyclonic loop originated at Cape Ghir
and the related surfacing of the poleward undercurrent as it flows north without opposition.
Fig. 8 is a schematic representation of the main
geostrophic circulation in this region, as described
above, from the mooring data and 3 years of XBT
transects in the region. Fig. 8a shows the situation
predominant most of the year and Fig. 8b represents
the large loop that may form in late fall. The exact
location of this loop, and how much of it flows north
along the upwelling zone, is indeed speculative and
needs to be verified through future measurements in
the region. Note that these schematic diagrams do not
include export of surface water through Ekman
transport or upwelling filaments. The main modifications produced by the filaments cell would be
localised distortions such as small loops, while the
vertical upwelling cell would cause an offshore drift
of the streamlines, mainly south of the Canary
Archipelago.
4.2. Dissolved inorganic carbon and nutrient fluxes
In November 1995 the Canary Islands area acted as
a source of CO2, with the maximum CO2 fluxes
occurring in the northwest coast of Fuerteventura,
where localised upwelling was taking place (SantanaCasiano et al., 2001). The effect of upwelling on the
CO2 system may also be observed in studies done
during October 1999 between the Canary Islands and
Cape Ghir. The introduction of subsurface water, rich
in nutrients and with relatively high fCO2 values, to
the nutrient-depleted ocean surface favours primary
production, and in turn reduces the amount of
dissolved inorganic carbon (CT) through photosynthesis. Using the above measurements, we may
estimate a decrease of about 26 Amol kg 1 in the
surface waters between the upwelling and open ocean
regions (from 2114 Amol kg 1 at station 615 to 2088
19
Amol kg 1 at station 619), approximately following
the trajectory of surface water parcels in this region
(Fig. 9a).
Phytoplankton uptake in the highly productive
upwelling waters moving offshore explains the
normalized partial pressure of carbon dioxide
(NfCO2) distribution in Fig. 9c, with maximum values
above 400 Aatm in the upwelled waters at the
northeastern extreme of the region. However, it is
the temperature field in the oligotrophic region that
controls the fCO2 values (Fig. 9b), with the offshore
area acting during this time of the year as a source of
CO2 to the atmosphere with an average value of 5
mmol C m 2 y 1 (González-Dávila et al., 2003).
The vertical circulation cell associated to coastal
upwelling transports coastal properties towards the
open ocean. Wooster et al. (1976) and Nykjaer and
Van Camp (1994) have obtained characteristic
values of cross-shore Ekman transport between the
Canary Island and the Strait of Gibraltar between 0
and 1.5 m2 s 1 (per meter of coastline), depending
on the season. If we multiply this figure by 5 mmol
m 3 (a typical mean value of the nitrate concentration within this layer in the upwelling band;
Minas et al., 1982), we obtain that the maximum
offshore nitrate transport is about 0.7 mmol s 1.
Over a distance of 1000 km this would account for
a net nitrate export up to some 7 kmol s 1 during
the maximum upwelling season. This value may be
compared with the amount of nitrate transported
alongshore by the Canary Current imbedded in the
upwelling system. Assuming a depth integrated
mean nitrate concentration of 8 mmol m 3 in the
top 400 m (Minas et al., 1982) and multiplying this
value by a southward water transport of 2 Sv leads
to a southward nitrate transport of some 16 kmol
s 1. Despite the involved approximations and even
taking into account the possibility that a significant
fraction of nutrients is lost at the filaments (acting
as point sinks), it is clear that an important fraction
of the nutrient finds its way south beyond the
Canary Islands. These results may be compared with
calculations of southward alongshore nitrate transport at about 328N (north of Cape Blanc) by
Marrero and Pelegrı́ (1995), between 5 and 7 kmol
s 1, which indicates the meridional continuity of the
southward nutrient transport in the upwelling system
off northwest Africa.
20
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Fig. 8. Schematic diagram illustrating the mean geostrophic transport of upper-thermocline waters (down to about 700 m) in the region. Each
line roughly corresponds to one Sverdrup. (a) Predominant situation during most of the year; (b) situation in late fall.
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
21
Fig. 9. (a) Temperature, (b) f CO2, and (c) Nf CO2 distributions in the upwelling transition zone north of the Canary Archipelago during October
1999. The location of stations mentioned in the text is shown in a; see also Fig. 4.
22
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
4.3. Plankton productivity in upwelling waters
Coastal upwelling off northwest Africa is characterised by high wind stress that forces relatively deep
mixing (Longhurst, 1998), such that nutrient differences in the upwelling source water set the potential
of primary production. Nutrient-rich South Atlantic
Central Water (SACW) upwells south of Cape Blanc,
with some interleaving across the meridional front
(Barton, 1987), while north of Cape Blanc the
upwelling water is typically NACW. Typical nutrient
contrast between NACW and SACW (at a level
immediately below the seasonal thermocline,
r uc26.5) is rather large, for nitrate decreasing
northwards from over 20 to 5 mmol m 3 (Zenk et
al., 1991; Pérez-Rodrı́guez et al., 2001). Compiled
productivity values in the whole region (Table 2)
yields an average annual estimate of 2.4 g C m 2
day 1, a value close to that computed from satellite
chlorophyll and averaged photosynthetic parameters
(Longhurst et al., 1995). Nevertheless, the highest
annual rates are probably produced between Cape
Barbas and Cape Blanc, where SACW is available as
source water and upwelling is produced all year long.
Significant changes in the nutrient regime and
primary production are observed on short-term scales,
correlated with variations in the local winds (Grall et
al., 1982). The centre of the upwelled water migrates
from the inner shelf to the shelf break as upwelling
progresses. The high wind stress and strong equatorward and cross-shelf currents may prevent accumu-
lation of organic matter in sediments, where aerobic
respiration dominates. Concentrations of particulate
organic matter nearshore are in the low end among all
coastal upwelling regions, with a predominance of
b50-Am particles (Lenz, 1982). However, mineralization rates and nutrient fluxes are as high as in other
upwelling regions (Minas et al., 1982). Fluxes of
nutrients from sediments into the well-mixed water
column are appreciable, and silica dissolution may
exceed rates of silica uptake by phytoplankton
(Codispoti et al., 1982). High ammonia concentrations
are homogeneously distributed in the turbid inshore
waters, where ammonia uptake by phytoplankton is
inhibited. Turbidity results from wind mixing and
aeolian dust deposition blown into the sea from the
Sahara desert. High nutrient regeneration rates may
result also from organic matter recirculated by
mesoscale features into the nearshore environment.
Fish populations contribute mainly to regeneration of
phosphate and urea, while ammonia is equally
regenerated both by zooplankton and fish (Smith
and Whitledge, 1982).
Organic carbon flux rates near the shelf bottom
may exceed primary production rates, indicating that a
large fraction of the nearshore material is resuspended.
This material may be lost out from the system by
cross-shelf advective transport (Longhurst, 1998).
Large filaments of cool water like those stretching
from Cape Ghir and Cape Blanc (Van Camp et al.,
1991) may export large proportions of coastal
produced organic matter hundreds of kilometers into
Table 2
Daily primary production values in coastal waters from the Northwest Africa upwelling system
Region
C.Sim–C.Ghir
C.Sim–C.Ghir
C.Sim–C.Ghir
C.Sim–C.Ghir
C.Jubi–C.Bojador
C.Jubi–C.Bojador
C.Corveiro–C.Blanc
C.Corveiro–C.Blanc
C. Blanc
C. Blanc
C. Blanc
C.Timiris–Nouakchott
NW Africa upwelling
Latitude (N)
31.5–30.5
31.5–30.5
31.5–30.5
31.5–30.5
28.5–26.5
28.5–26.5
22.0–21.0
22.0–21.0
21.0
21.0
21.0
19.5–18.0
31.5–18.0
Season
Summer
Winter
Summer
Autumn
Summer
Summer
Spring
Spring
Spring
Spring
Summer
Spring
Daily primary production (g C m
2
day 1)
Average
Min.
Max.
2.4
1.3
1.3
1.5
3.1
2.4
2.4
0
0.2
0.6
1.0
1.3
4.2
2.5
2.8
2.5
5.3
0.8
1.1
0.3
0.2
0.4
1.6
5.0
3.4
2.6
1.8
3.6
4.3
1.15
0.96
1.62
3.9
2.4F1.5
Reference
Minas et al., 1982
Minas et al., 1982
Grall et al., 1982
Head et al., 1996
Basterretxea and Arı́stegui, 2000
Minas et al., 1982
Huntsman and Barber, 1977
Lloyd, 1969
Schulz, 1982
Schulz, 1982
Schulz, 1982
Minas et al., 1982
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
the ocean domain (Gabric et al., 1993; Garcı́a-Muñoz
et al., 2005). Mineralization of organic matter along
filaments must constitute an important nutrient source
for new production in nearby oceanic communities.
Zooplankton biomass in the upwelling area varies
with latitude and time, in relation with changes in
wind intensity (Postel, 1990). The highest values are
found near Cape Blanc along the entire annual cycle.
To the north the maximum values are observed in
summer and the minimum in winter, the opposite
situation occurring south of Cape Blanc. The increase
in zooplankton biomass appears to be uncoupled with
the maximum of phytoplankton biomass and changes
in zooplankton composition and growth are observed
as the water mass is transported offshore. The species
composition varies depending on the seasonal upwelling cycle of the upwelling (Thiriot, 1978). Calanoides
carinatus is the most common species during the
upwelling season. During the non-upwelling season it
remains in overwintering at mesopelagic depth (Vives,
1975), while during the upwelling season it appears at
the surface where it grazes the high diatom densities
usually found over the African shelf. Maximum
abundance is found near the upwelling centres
coinciding with high values of primary production.
As the water masses are transported offshore, C.
carinatus is observed at increasing depth coinciding
with the decrease in primary production and phytoplankton biomass. Euphausiids are also observed
offshore as strong diel vertical migrants. Euphausia
khronii is the most abundant species (80–90% of the
euphausiids) north of Cape Blanc. Other organisms
such as Thalia democratica are observed over the
slope waters forming high-density swarms.
5. Upwelling filaments
5.1. Structure and dynamics
Considerable knowledge on the dynamics and
structure of filaments came after the Coastal Transition Zone interdisciplinary program off California.
A summary of the program may be found in Brink and
Cowles (1991) and a description of the major physical
characteristics of the observed filaments was given by
Strub et al. (1991). Coastal filaments are relatively
narrow (typical widths of a few tens of kilometers)
23
near surface (usually no deeper than 150 m) structures
that stretch offshore (up to several hundred kilometers) from near the coast. The core of a filament
flows offshore with rather large velocities, usually
between 0.25 and 0.5 m s 1, and it is common to find
onshore flow at both sides of the filament associated
to one or two vortices. These structures have been
observed in all eastern boundary currents of subtropical gyres and are almost always associated to
upwelling favourable conditions, such that they
contain relatively cold water of subsurface origin.
According to Strub et al. (1991) and Ramp et al.
(1991), filaments may arise from one or a combination of several factors: baroclinic instability of the
coastal current, irregularities in coastline and bottom
topography, coastal convergence caused by wind
stress, and the interaction of the offshore eddies with
the coastal region. One important element to consider
is that filaments are a very common feature off
California during the upwelling season, when the
California Current flows south. During winter the
winds reverse, upwelling ceases and the current (now
the Davidson Current) flows north, filaments being
very uncommon. Other clues on the origin of
filaments are that they contain nutrient-rich cold water
of subsurface origin and that their location sometimes
appears to be related to major coastal or topographical
features, but in other instances there is no clear
association.
A similar situation occurs at the eastern boundary
of the North Atlantic subtropical gyre. Off the Iberian
Peninsula, where the alongshore current reverses
seasonally, filaments are a common feature during
the upwelling season (Haynes et al., 1993) but they
are absent when the water flows north (Haynes and
Barton, 1990; Barton, 1998). South of the Strait of
Gibraltar upwelling takes place all year long, although
with a clear summer maximum, and the Canary
Current continuously flows south. Filaments are here
a recurrent feature all year long, although the small
winter surface temperature contrast between coastal
and open waters makes it hard to appreciate their SST
signal during this season.
In any satellite SST image, we usually appreciate a
number of filaments emanating from the northwest
African coast, two of them being major permanent
structures: the Cape Ghir and the Cape Blanc
filaments. The latter is a giant filament that typically
24
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
stretches nearly 500 km off Cape Blanc. It lies south
of our region of interest, its great size and persistence
being linked to the convergence of North Atlantic and
South Atlantic upper thermocline waters at the Cape
Verde frontal zone (Gabric et al., 1993). The Cape
Ghir filament is a rather ubiquitous feature that
usually stretches offshore over 150 km. It is visible
in most colour and SST satellite images of the region
(e.g. Plate 1; see also Van Camp et al., 1991 and
Hernández-Guerra et al., 1993) and even appreciable
from the mean surface temperature fields, obtained
from either hydrographic (Mittelstaedt, 1991) or
satellite data (Hernández-Guerra and Nykjaer, 1997).
Several hydrographic cruises off Cape Ghir have been
documented by Hagen et al. (1996) and Pelegrı́ et al.
(2005). These cruises covered the Cape Ghir plateau,
off the northern flank of Cape Ghir, and the Agadir
submarine canyon, immediately south of this Cape.
The filament was always present, even during nonfavourable upwelling winds, and consisted in a
shallow feature that did not extend beyond about
200 m. Water flows offshore in the central portion of
the filament, with characteristic speeds over 0.25 m
s 1, and recirculates onshore at both sides of the
filament as two oppositely rotating eddies. South of
the filament a deep anticyclonic eddy, its origin
probably linked to the poleward undercurrent over
the slope, was found in two October cruises. Fig. 10a,
reproduced from Pelegrı́ et al. (2005), presents the
temperature field at a depth of 20 m and the velocity
field obtained from ADCP measurements in the 25- to
75-m layer at a time of a relatively weak filament that
resulted from non-favourable upwelling winds.
South of Cape Ghir, it is usual to find other
filaments emanating from the coast. HernándezGuerra et al. (1993) showed different satellite images
with filaments either detaching from Cape Yubi or
Cape Bojador, and Barton (1998) and Barton et al.
(1998) discussed a situation when a large filament
emanated from somewhere between these two Capes.
There are many instances, however, when there are no
apparent filaments in this region (Nykjaer, 1988; Van
Camp et al., 1991; Arı́stegui et al., 1994). Other minor
filaments have also been observed in the northwest
coast of Fuerteventura in the Canary Islands (Real et
al., 1981; Molina and Laatzen, 1989).
All the above features point at the relation between
filaments and coastal upwelling, and suggest that
under some circumstances they may indeed be linked
to the baroclinic instability of an intense coastal
current, probably triggered by topographic features.
The reasoning behind this argument is that during the
upwelling season the coastal jet is intense enough
such that it becomes unstable. The potential energy
available in the eastern boundary system would justify
that filaments appear even when the upwelling
favourable conditions weaken. This would also justify
that no filaments are present when the eastern
boundary flows north, since these flows have
downwelling isopycnals at the coast and no potential
energy is available from this system. It seems
surprising, however, that filaments are absent in
intense western boundary currents. This suggests that
an additional mechanism, based on the importance of
bottom friction acting over steep slopes, could also be
relevant. Lee et al. (2001) have proved that the main
source of relative vorticity over the continental slope
is the bottom stress acting over the seaward deepening
water column. In a southward current flowing over the
continental slope, this mechanism would provide
positive vorticity so the flow would not be capable
to adjust to the local planetary vorticity. When this
happens we could expect that it would move away
from the slope. This mechanism relies on the
existence of a bottom boundary layer over the
continental slope but this has indeed been observed
to be important in the region (Badan-Dangon et al.,
1986).
5.2. Distribution of nutrients and carbon
Several authors (Minas et al., 1982; Basterretxea,
1994; Garcı́a-Muñoz et al., 2005) have reported the
existence of relatively high levels of nutrients at the
base of a filament, in concordance with the subsurface
origin of these waters. Nutrient concentrations
decrease in the offshore direction and become rather
exhausted at the offshore extreme of the filament. The
nutrient field is coherent with the temperature field
and maintains an inverse linear relation with chlorophyll vales. These characteristics indicate that
nutrients are being utilised as the water parcels move
offshore, and suggest that the direct export of nutrients
by the filaments to the open ocean may be relatively
small, at least compared with the export of organic
mater. The regeneration of exported organic matter,
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
25
Fig. 10. Distribution of properties in the Cape Ghir region from September 28 to October 1, 1997. (a) Temperature field at 20-m depth (left) and
velocity vectors in the 25- to 75-m layer (right) adapted from Pelegrı́ et al. (2005). (b) Mean depth (top 100 m) distribution of Chlorophyll a
(Chl a, mg m 2), particulate organic carbon (POC, mmol m 2), dissolved organic carbon (DOC, mmol m 2) and nitrate+nitrite (mmol m 2).
Reproduced from Garcı́a-Muñoz et al. (2005).
26
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
however, causes an indirect export of both nutrients
and dissolved organic carbon, reflected in a high
correlation between these quantities (Minas et al.,
1982; Pérez et al., 2001). Fig. 10b, adapted from
Garcı́a-Muñoz et al. (2005), illustrates some of these
features at a time of a relatively weak filament.
Despite most nutrients appear to be utilised along
the offshore excursion of water parcels, it is
unquestionable that filaments represent a point
source of nutrients along the coast of magnitude
much larger than the source of nutrients. Garcı́aMuñoz et al. (2005) have computed a net export of
nitrate+nitrite of 2.6 kmol s 1 by the Cape Ghir
filament, most of it related to significant nutrient
concentrations and offshore velocities at depths
between 50 and 100 m. This value is indeed much
larger than the offshore export that would take place
in an upwelling system along a shore distance of
only a few tens of kilometers, equivalent to the
filament width. Using a maximum water Ekman
transport of 1.5 m2 s 1 (per meter of coastline) and
a typical value of 3 mmol m 3 for nitrate
concentration at the base of the seasonal thermocline, we may estimate that the nitrate Ekman
transport through a 20-km wide filament would be
about 0.1 kmol s 1, less than 5% the actual
advective transport estimated by Garcı́a-Muñoz et
al. (2005). This clearly points at the character of
filaments as locations of coastal flux convergence
and export to the open ocean.
The effect of upwelling filaments on the CO2
system may also be observed from the measurements
taken in the area near Cape Ghir during October 1999
(Fig. 9). Filaments stretch into the open ocean and
introduce a perturbation in the distribution of chemical characteristics, the most visible feature being a
change in the partial pressure of CO2 in surface
waters. We may use stations 615 (within the Cape
Ghir filament) and 612 (just out of the filament) to
visualise what happens to the carbon content of a
water parcel exported away of the upwelling region
through a filament. Total dissolved inorganic carbon
and partial pressure of CO2 increase by 13 Amol kg 1
and 19 Aatm, respectively (Fig. 9b). These values,
however, cannot be directly used to conclude whether
the region acts as a source or sink of CO2. Before
doing so, we must consider the thermodynamic
(temperature and salinity change) and air–sea
exchange effects. Our observations indicate that the
surface temperature, salinity and pH increase from the
upwelling region to the open ocean. Temperature
increases by 3.9 8C and salinity and pH increase by
about 0.2 and 0.05 units, respectively. Normalisation
to a constant temperature of 21.5 8C (NfCO2)
indicates that the f CO2 experiments a large increase
in the upwelling region because of the sole thermodynamic effect in CO2 solubility (Fig. 9c). This is
consistent with the fact that in the upwelling region
there is upward pumping of subsurface CO2 rich
waters, i.e. f CO2 augments by some 40 Aatm at a
depth of 100 m.
For a water parcel moving from station 615 to
station 612, for example, the thermodynamic effect
would increase f CO2 by 64 Aatm while the air–sea
exchange effect on the mixing layer would produce an
increase of less than 1 Aatm. Hence, in the absence of
biological effects, we would expect that the water
parcel experimented changes much larger than
actually observed. This suggests that there is significant nutrient utilisation and biological production in
the area. For example, to consume 28 Amol kg 1 of
dissolved inorganic carbon (corresponding to a
change of 45 Aatm in f CO2, or the difference between
the thermodynamic imbalance and the actual increase
between our reference stations) requires 3.9 Amol
kg 1 of nitrate (CORG:N=7.2F0.8, Körtsinger et al.,
2001). The time scale for utilization of this nitrate is
relatively low, typically of the order of a few days
(Louanchi and Najjar, 2001, Agustı́ et al., 2001).
Using an outward velocity value of 0.1 m s 1 and
considering that the two stations are separated by a
distance of 73.5 km, we may estimate a utilization
time of 8.5 days. For a 70 m deep euphotic layer, this
corresponds to a mean nitrate utilization rate of about
1 mmol m 3 day 1. This estimate is of the same order
as the Agustı́ et al. (2001) calculations for the
downward average flux of dissolved organic nitrogen
in the eastern subtropical Atlantic, 15 mmol m 2
day 1, pointing at coastal filaments as a major input
of organic nitrogen and carbon to this region.
The mean CT over the upper 150 m decreases from
2114 Amol kg 1 at station 615 to 2100 Amol kg 1 at
station 612. This decrease in such a small distance
suggests that the biological utilization of the injected
nutrients inside the filaments decreases the concentration of dissolved inorganic carbon. The upward and
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
offshore supply of inorganic nutrients, however, does
not necessarily lead to a net uptake of atmospheric
carbon, since this nutrient supply is associated with a
supply of dissolved inorganic carbon. It is the low
temperature of the upwelled water and the concentration of inorganic carbon in the source water that
determine whether the upwelling area will behave as a
net CO 2 sink or source. At the time of our
observations, during an intense upwelling event, we
may conclude that physico-chemical processes dominated over the biological pump in determining that the
upwelling region behaved as a net CO2 sink. These
results are confirmed by observations of an average
carbon flux of 0.5 mmol m 2 day 1 from the
atmosphere to the ocean in the upwelling region
(González-Dávila et al., 2003). A similar finding was
described by Agustı́ et al. (2001) around this area with
a net upward flux of dissolved inorganic carbon and
downward flux of organic carbon.
5.3. Plankton distribution and organic matter
transport
Upwelling filaments may enhance the exchange of
biological properties between upwelling and open
ocean waters, since their offshore transport is usually
significantly larger than Ekman transport (Kostianoy
and Zatsepin, 1996; Navarro-Pérez and Barton, 1998).
Nevertheless, the return flow of filament water into
the coastal upwelling jet may reduce the impact of the
overall offshore advective transport. In August 1993
an upwelling filament stretching between Cape Jubi
and Cape Bojador was surveyed. The filament
extended approximately 150 km offshore where it
wrapped around a cyclonic eddy of 100-km diameter,
before returning shoreward. Waters with high chlorophyll content, dominated by phytoplankton cells
larger than 2 Am, were mainly restricted to the shelf
and to a few stations along the filament. Although
grazing could actively control the phytoplankton
abundance along the filament (Hernández-León et
al., 2002), the rapid warming of surface water with
distance offshore probably induces mass sedimentation of diatoms in slope waters, and a drastic change
in the phytoplankton community structure (e.g. Brink
and Cowles, 1991). Indeed, a sharp drop in chlorophyll ascribed to cells larger than 2 Am was evident
in the boundary between the upwelling and filament
27
waters (Basterretxea and Arı́stegui, 2000), coinciding
with a change from a diatom-dominated community in
onshore waters to a cyanobacteria-dominated community in filament waters (Van Lenning, 2000). This
scenario is similar to the one described by Joint et al.
(2001), who observed drastic changes in plankton
community structure within a filament of the Iberian
upwelling system as it progressed offshore. Near the
coast they found high chlorophyll concentrations
associated with a diatom-dominated assemblage,
which included high biomass of heterotrophic dinoflagellates and ciliates, while in the offshore extension
of the filament they observed low chlorophyll values
associated with a picoplankton community.
Basterretxea and Arı́stegui (2000) have found that
phytoplankton assimilation numbers, and thus primary productivity, show a marked gradient from
upwelling to open ocean waters. The assimilation
numbers changed over one order of magnitude,
paralleling changes in biomass and phytoplankton
community structure. During their study they found
that the filament was clearly acting as a recirculation
loop. Seaward biomass transport occurred mainly in
the mixed layer, whereas shoreward transport presented maxima at deeper layers, the net transport
being onshore. This situation, however, differs from
other field studies (Arı́stegui et al., 1997; Garcı́aMuñoz et al., 2005; Arı́stegui and Montero, 2005), or
from studies based on ocean colour images (e.g.
Hernández-Guerra et al., 1993), in the same region,
where high organic matter concentrations transported
by filaments were observed. Arı́stegui and Montero
(2005) showed that coastal upwelling waters invading
the eastern Canary region in the form of filaments can
transport either water with low plankton respiration
and large phytoplankton cells or water with high
respiratory rates associated with small cells. Hence,
the nature of the filament and its genesis, as well as
the stage of development of the filament structure, are
critical for the biological consequences of the filament
in terms of plankton productivity and export of carbon
to the open ocean.
A close relationship between the structure of the
filament and zooplankton biomass distribution has
been identified by Hernández-León et al. (2002). They
find high biomass, specific gut fluorescence and
electron transfer activity (ETS) along the filament
structure in contrast to relatively low values of these
28
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
parameters at the centre of a cyclonic eddy enclosed
by the upwelling filament. High values of zooplankton biomass are observed in all the filament structure
in contrast to chlorophyll and primary production,
which decrease towards the African shelf. The
longevity of these organisms as well as a change
from a plant-based to a non-pigmented (protista) diet
allows these organisms to move up to 100 km from
the coast. Of great interest is the transport of fish
larvae observed in the filament structure (Rodrı́guez et
al., 1999). Neritic species such Sardina pilchardus are
found along the filament signature, while oceanic
species are excluded. These neritic species are good
tracers of the filament physical structure, which
transports them into oceanic waters and near the
Canary Islands (Bécognée et al., unpublished manuscript), where they have an important impact in the
local fisheries.
6. Island-generated mesoscale structures
6.1. Generation and dynamics
The Canary Archipelago acts as a barrier to the
prevailing winds and currents, representing an exceptional natural example of flow perturbation by
obstacles. Two main features arise as a consequence
of flow perturbation: cyclonic and anticyclonic eddies
and warm lee wakes. Cyclonic eddies, detaching from
the western flank of the islands, have been identified
downstream of the islands of La Palma, Gomera and
Tenerife from satellite observations (HernándezGuerra et al., 1993; Pacheco and Hernández-Guerra,
1999; Barton et al., 2000), and downstream of Gran
Canaria island from both satellite and in situ
observations (Arı́stegui et al., 1994, 1997; Hernández-Guerra et al., 1993; Pacheco and HernándezGuerra, 1999; Barton et al., 1998, 2000; Marrero-Dı́az
et al., 2001; Basterretxea et al., 2002). Anticyclonic
eddies, detaching from the eastern flank of the islands,
have been observed downstream of Tenerife and Gran
Canaria islands from both satellite and in situ
observations (Arı́stegui et al., 1994, 1997; Pacheco
and Hernández-Guerra, 1999; Barton et al., 1998,
2000, 2001; Molina et al., 1996). SST images have
also shown anomalous high surface temperature
filaments that stretch southwest at the lee side of
most archipelago islands, in the direction of the mean
velocity surface field (Hernández-Guerra, 1990; Van
Camp et al., 1991). These warm oceanic wakes, or lee
filaments, are surface structures whose origin lies in
the intense day-time warming of the unstirred surface
in the wind-sheltered area of the islands. They occur,
to some extent, in all the islands of the archipelago but
are particularly clear downwind of the higher islands:
La Palma, Gomera, El Hierro, Tenerife and Gran
Canaria (Barton et al., 2000, 2001; Basterretxea et al.,
2002).
Eddies are a recurrent mesoscale feature related to
the Canary Current perturbation by Gran Canaria.
Pacheco and Hernández-Guerra (1999) analyzed 8
years of satellite colour (Coastal Zone Colour
Scanner, CZCS) images and observed the signal of
cyclonic eddies associated with Gran Canaria,
Gomera, and La Palma and anticyclonic eddies
associated with Gran Canaria all during the year.
Eddies have also been detected in most surveys
conducted south of Gran Canaria, capable of resolving
mesoscalar structures (Arı́stegui et al., 1994, 1997;
Barton et al., 1998; Basterretxea et al., 2002). Table 3
summarises the number of eddies sampled during
different surveys carried out on Gran Canaria waters,
with eddies observed during all seasons. Cyclonic
eddies were located southwest of Gran Canaria, either
attached to the coast or at approximately one island
diameter (50 km) from it. Anticyclonic eddies have
been located either attached to the south coast or
approximately at one island diameter from it. The
high spatial coverage provided by satellite images
indicates that the simultaneously occurrence of
cyclonic and anticyclonic eddies downstream of Gran
Canaria, forming a von Kármán like vortex street, is a
quite common feature.
Table 3
Summary of eddies sampled during different surveys carried on
south of Gran Canaria Island
Survey
Date
Cyclonic Anticyclonic
AXBT-1
AXBT-2
AXBT-3
EMIAC 9006
EMIAC 9103
CANARIAS 9110
MAST 9308
FRENTES 9806
5 May, 1989
20 February, 1990
24 May, 1990
6–12 June, 1990
9–17 March, 1991
25 Oct–5 Nov, 1991
18–24 Agust, 1993
6 June–8 July, 1998
1
1
1
1
2
1
1
1
1
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
The mean radius of eddies generated by an island is
comparable to the island size. These eddies usually
exhibit an elliptical shape, in a similar fashion as
observed by Patzert (1969) for eddies shed from
Hawaii Island. Patzert suggested that the elliptical
shape indicates that the eddy is at an early stage, and
argued that an eddy’s eccentricity decreases with
increasing distance of the island. Pacheco and
Hernández-Guerra (1999) used CZCS images to
calculate the eccentricity as a function of distance
from the island, for an anticyclonic eddy associated
with Gran Canaria and for a cyclonic eddy associated with Gomera. The results show that the Gran
Canaria anticyclonic eddy eccentricity decreases
when the eddy moves further away from the island.
However, the Gomera cyclonic eddy behaves in a
completely inverse way. They explained this anomalous behaviour, pointing out that the vicinity of El
Hierro Island may disturb the normal evolution of
the eddy.
As a consequence of geostrophic adjustment, the
cyclonic eddies have a central uplifting of the
isotherms, being cold core water structures, whereas
anticyclonic eddies cause the central downlifting of
the isotherms, becoming warm water core structures.
The perturbation of the temperature field by an eddy is
noticeable from the near surface layer until at least 600
m (Sangrà, 1995). Intense cyclonic eddies may uplift
29
the isotherms by some 60 m, introducing temperature
negative anomalies at about 2.5 8C. On the other hand,
the sampled anticyclonic eddies seem to be less
intense, causing a depression of the isotherms of the
order of 45 m and introducing positive temperature
anomalies about 1 8C. Secondary ageostrophic circulation associated to cyclonic (anticyclonic) eddies
appears to cause upwelling (downwelling) of deep
(surface), nutrient-rich (-poor), waters in their cores as
a results of outward (inward) advective flux of surface
water. From the outward trajectory of an Argos buoy,
deployed in a cyclonic eddy shed by Gran Canaria,
Barton et al. (1998) estimated an upwelling velocity
higher than 50 m day 1.
Three Argos buoys where deployed on an anticyclonic eddy spun off by Gran Canaria in early June
1998 (Sangrà et al., 2004). Fig. 11 shows the whole
trajectory of one buoy, an anticyclonic eddy-like path
lasting during 199 days, which suggests that Gran
Canaria eddies could last as a coherent structures for
many months. The rotation period of all three buoys
increased from 3 to 6 days, which indicates that the
rate of rotation of the eddy decreases along its
lifetime. The initial radius of the eddy, derived from
an XBT survey conducted previous to the buoy
deployment, was about 25 km, while the final radius,
computed from the buoy’s orbit, was about 50 km.
This suggests that the radius of the eddy also increases
Fig. 11. Track of an Argos buoy, dragged at a depth of 100 m, following an anticyclonic eddy shed by Gran Canaria Island. Dots are plotted at
10-day intervals and annotated every 20 days. Day 0 corresponds to June 30, 1998, and day 199 to January 14, 1999.
30
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
along its life time. The orbital trajectories (computed
subtracting the mean eddy translation) also show an
oscillation of the radius around its mean value, which
suggests that the eddy evolves on a pulsating mode.
One plausible mechanism for the generation of
eddies is the perturbation of the Canary Current by the
islands, but this depends on whether the Canary
Current is energetic enough to produce vortex
shedding. Sangrà (1995) demonstrated numerically
that if the Canary Current remains above 0.1 m s 1
then a von Kármán street is formed downstream of
Gran Canaria. Direct measurements and geostrophic
calculations indicate that a mean velocity of 0.2 m s 1
may occasionally be attained (Sangrà, 1995) but,
according to Navarro-Pérez and Barton (2001), the
maximum mean geostrophic value of the Canary
Current through the Canary Archipelago is relatively
low, roughly 0.05 m s 1. Therefore, only during
certain times of the year the Canary Current is
energetic enough to generate eddies. During low
current velocity and moderate trade wind periods,
another mechanism that favours eddy generation may
be Ekman pumping originated through wind shear at
the islands wakes (Barton et al., 2000, 2001;
Basterretxea et al., 2002). Recent numerical simulation of eddy shedding by Gran Canaria, which
includes wind forcing, shows that the vorticity input
by the wind through Ekman pumping at the island
wake helps to produce vortex shedding at relatively
low incident current velocity (Jiménez and Sangrà,
1999). For example, for a wind of 10 m s 1 vortex
shedding is produced when the incoming current is
above just 0.04 m s 1, which approaches the mean
values reported by Navarro-Pérez and Barton (2001).
6.2. Source of nutrients and dissolved inorganic
carbon
Eddies generated at the lee of the Canary Islands
produce an important perturbation in the distribution
of chemical properties in the surface euphotic layers.
Cyclonic eddies are a source of subsurface nutrientrich water into the euphotic layers (Arı́stegui et al.,
1994, 1997). There are several mechanisms through
which this may occur. First, the central core of
cyclonic eddies in geostrophic balance is characterised
by the upwelling of isopycnals, hence physicallly
bringing the subsurface layers close to the surface
(Section 6.1). Second, there appears to be actual
ageostrophic isopycnal flow into the eddy core,
particularly during the early stages of eddy development when geostrophic equilibrium has not yet being
fully attained, which leads to near surface central
divergence (Barton et al., 1998). And third, diapycnal
mixing likely takes place at different eddy phases and
locations, probably linked to a localised increase in
diapycnal shear (Rodrı́guez-Santana et al., 1999,
2001). The combination of these mechanisms is
probably what makes these cyclonic eddies to be an
efficient source of nutrients for the surface euphotic
layers. Water parcels in early stages of anticyclonic
eddies, on the other hand, appear to undergo radial
motions towards the eddy centre of order 0.01 m/s, as
suggested from the analysis of surface buoy trajectories (Fig. 11). This radial convergence leads to surface
central waters low in nutrient concentration but with
high accumulation of biomass.
Although cyclonic and anticyclonic eddies have a
quite opposite behaviour, which leads to characteristic
different nutrient and chlorophyll distributions, they
both appear to undergo intense diapycnal mixing at
intermediate distances between their centre and
periphery, where diapycnal shear is maximum. Vertical (diapycnal) velocities in these regions may be as
large as 10 5 m s 1 (A. Rodrı́guez-Santana, personal
communication), which would explain the intermittent
character of nutrient anomalies observed at some
stations (Arı́stegui et al., 1997).
The area between La Palma, La Gomera and
Tenerife is characterised by coexisting cold- and
warm-core eddies, which contribute to exchanging
CO 2 between the atmosphere and the ocean
(González-Dávila et al., 2005). The cold-core
cyclonic eddies have surface waters with high
partial pressure of CO2 and nutrient concentrations
inside the eddy, while relatively low partial pressure
of CO2 and nutrient values are observed near the
eddy limits. Warm-core anticyclonic eddies are
associated with low partial pressure of CO2 and
relatively high chlorophyll concentration inside the
eddy. Surface values of fCO2 show that this area
acts as a source of CO2 to the atmosphere, but with
a value that is only half (5.4F2.0 mmol m 2 day 1)
of that determined at the same time north of the
Canary Islands and during similar wind conditions
(González-Dávila et al., 2005). This is due to the
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
effects of both the solubility and the biological
pumps associated with the cyclonic eddies where
chlorophyll enhancement follows a decrease in CT.
6.3. Plankton response to mesoscale variability
Nutrient pumping into the euphotic zone by
cyclonic eddies enhances phytoplankton primary
production and consequently increases chlorophyll
concentrations (Arı́stegui et al., 1997; Basterretxea,
1994; Barton et al., 1998; Basterretxea and Arı́stegui,
2000). Conversely, anticyclonic eddies concentrate
warm surface water and organic matter in their cores,
deepening the mixed layer and sinking organic matter
below the euphotic zone (Arı́stegui et al., 1997;
Arı́stegui and Montero, 2005). Counter-paired
cyclonic and anticyclonic eddies act therefore as a
two-way biological pump, increasing plankton production in cyclonic eddies and accelerating the
vertical transport of organic matter out of the photic
zone within anticyclonic eddies (Arı́stegui and Montero, 2005).
Anticyclonic eddies associated with Gran Canaria
are revealed by chlorophyll-like pigments (CZCS)
images, thanks to their interaction with upwelling
filaments stretching from the African coast (Hernández-Guerra et al., 1993; Pacheco and HernándezGuerra, 1999). Upwelling filaments export water with
relatively low temperature and high chlorophyll
content that is entrained by the anticylonic eddies
drawing its contours. Therefore, anticylonic eddies
may play a major role on the export of highchlorophyll upwelling waters into the interior oligotrophic ocean. Satellite images and in situ data also
indicate that high-chlorophyll water, resulting from
island stirring or local upwelling in the western flanks
of the islands, is incorporated into the cyclonic eddies
during their early stages and subsequently advected
downstream by the Canary Current (HernándezGuerra et al., 1993; Arı́stegui et al., 1997; Pacheco
and Hernández-Guerra, 1999).
Wind/current shear at the flanks of the islands may
enhance plankton metabolism through different
effects. Microplankton and zooplankton respiratory
activity is enhanced at coastal sites, where velocity
shear is maximum, presumably due to an increase in
turbulence (Hernández-León, 1988, 1991; Arı́stegui
and Montero, 2005). Ekman pumping on the wind
31
shear boundaries of the islands produces convergence
and divergence fronts which affect plankton distribution and productivity. Divergence fronts at the lee of
the island induce upwelling of thermocline water,
leading to an increase in primary production (Arı́stegui et al., 1989; Basterretxea et al., 2002). At these
fronts, zooplankton is accumulated either because of
the availability of large phytoplankton cells (Arı́stegui
et al., 1989) or through passive accumulation of
drifting zooplankton (Hernández-León, 1988, 1991;
Hernández-León et al., 2001a).
7. Conclusions
The surface and upper-thermocline waters of the
Canary Basin exemplify a southward flowing eastern boundary system, exhibiting features common to
these systems but also displaying several quite
specific characteristics. The dominating open ocean
current is the all year long southward flowing
Canary Current, the easternmost extension of the
Azores Current. Recent results have indicated that
the intensity and location of the Canary Current are
intimately linked to the coastal upwelling system off
northwest Africa, which is the true boundary
condition for the interior subtropical gyre. A striking
feature is that the Canary Current flows south in a
narrow band, typically no wider than 200 km,
linked to the coastal upwelling jet that develops
north of Cape Ghir. It is not yet clear how this
coastal current accommodates about 2 Sv of surface
and upper-thermocline waters inflowing from the
open ocean. A likely hypothesis is that this flux is
shared between the baroclinic jet, of relatively fast
response that increases to a maximum during
summer (the maximum upwelling season), and the
barotropic jet, of great inertia that keeps increasing
until late fall.
The baroclinic jet is a relatively shallow feature
with a width of the order of the internal Rossby
radius of deformation, about 100 km in the upwelling
region. This feature is present all year long but
intensifies in summer, during the maximum upwelling season, distinguishable in SST images as a frontal
region (even in the mean SST field, HernándezGuerra and Nykjaer, 1997). Coastal filaments, mainly
generated through baroclinic instability, are largest
32
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
and easiest to identify during this season. These
filaments export water and coastal characteristics into
the adjacent upwelling transition zone but it is likely
that a large fraction recirculate towards the coast. The
barotropic jet, on the other hand, probably reaches its
maximum intensity in late fall and becomes unstable
because of the inability of such a current flowing
south to adapt to the decreasing planetary vorticity.
At this time the Canary Current separates from the
coast and flows south through the Canary Archipelago, showing a meander-like large-scale offshore
diversion. When this happens the poleward undercurrent reaches the surface and northward flow may
be observed between the Canary Archipelago and the
African coast.
These features cause the upwelling transition zone
in the Canary Basin to be characterised by two cells
transporting subsurface upwelled waters into the open
ocean. The first one is the standard vertical cell,
present in all upwelling systems, with Ekman offshore
transport responding to the wind intermittent (synoptic and seasonal) variations. The second one is the
horizontal circulation cell that has its origin on open
waters impinging into the coastal region north of Cape
Ghir and is closed by water exported through several
filaments and the seasonal flow diversion at Cape
Ghir. Most of the year, the current system has
meridional continuity—thanks to the coastal upwelling system—and transports large amounts of nutrients
(order 10 kmol s 1 of nitrate and proportional
quantities of other nutrients) that sustain the fisheries
in the region (independent on the intermittency of
upwelling winds). The horizontal cell that detaches at
Cape Ghir, however, is a very peculiar seasonal
characteristic of the upwelling region in the Canary
Basin and provides a diversion of the nutrient transport into the open ocean. The joint action of both cells
causes this coastal upwelling region to be characterised by relatively low productivity values, as compared with other upwelling regions, and by a
substantial export of nutrients and organic matter into
the open ocean.
South of the Canary Archipelago, the Canary
Current is greatly perturbed because of vortices
generated at the islands and their interaction with a
number of coastal upwelling filaments. In contrast
with the Cape Ghir filament, the position of these
filaments is variable, which gives rise to the
hypothesis that their origin is linked to the
intensity of the southward flowing coastal current
and their interaction with the eddies. The detaching
vortices are a rather recurrent feature that appears
to be generated topographically by the impinging
current, although the northeasterly winds may help
their development. Vortices also are temporally
coherent structures, which have been followed for
a time period over 6 months while drifting some
500 km, so they must have a quite important
impact on the distribution of properties in the
whole eastern boundary system. Cyclonic (cold
core) vortices are characterised by the central
raising of the isotherms, and actual radial convergence at least during the early stages of their
development, while the opposite situation is true
for anticyclonic (warm core) vortices. Vortices also
appear to undergo intermittent diapycnal mixing, its
overall effect being the raising of nutrients to the
surface layers, which likely helps to maintain some
level of productivity at more mature stages of their
development.
From a biological point of view, the Canary
region behaves as a carbon source, both increasing
the concentration of CO2 in surface waters and
exporting organic carbon to the oceanic domain.
Compiled gross primary production (Pg) and community respiration (R) data from the region (Duarte
et al., 2001) indicate that R is several times higher
than Pg most of the year. This metabolic imbalance
can be only explained by allocthonous inputs of
organic matter from nearby regions. Coastal upwelling filaments, with high-chlorophyll concentration,
are frequently observed extending up to several
hundred kilometers offshore from the northwest
Africa coastal upwelling system. Eventually, island
eddies may interact with the offshore limit of some
of these filaments, entraining water with organic
matter and transporting it, through a complex eddy
field, to the oligotrophic open ocean. Although some
of the filament water may recirculate back to the
coastal jet, the regular analysis of SST and surface
chlorophyll from satellite images suggests that
upwelling filaments spread organic matter in a
recurrent fashion into the Canary region. A major
seasonal feature is the diversion of water from the
coastal upwelling region into the open ocean in the
Canary Archipelago region, during late fall and early
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
winter. Such diversion would be responsible for the
observations of high content of organic carbon
during this season in the open ocean, much larger
than actual production values (Arı́stegui et al., 1997).
Despite the high respiration rates measured in the
upwelling region, some of the organic matter is
indeed transported to the oligotrophic waters of the
subtropical gyre. Relatively high dissolved organic
carbon values have been measured throughout the
whole transition zone of the Canary region, south of
the archipelago, indicating that this semi-refractory
carbon is available to be advected out of the region
while it is slowly oxidized.
An additional source of organic carbon is due to
cyclonic eddies, which increase primary production
several times with respect to ambient waters, as a
response to nutrient pumping. Barton et al. (1998)
estimated that island eddies may contribute to the
nitrogen flux to the Canary region as much as
coastal upwelling. These eddies would have an
antagonist effect on the carbon fluxes in the region,
when cold, nutrient-rich and CO2-rich waters upwell
at their cores. The surface warming of the recently
upwelled water would produce an increase in surface
pCO2. Conversely, the biological fixation of CO2 by
phytoplankton would drop the CO2 surface concentrations. Overall, however, the net balance is
governed by the physical processes, and the cyclonic
eddies behave as net sources of CO2 to the
atmosphere. Finally, the fresh organic matter produced by eddies may either be partly oxidized in
short times (increasing R in surface waters of the
region), exported to the oligotrophic subtropical gyre
by the eddy field (increasing R in the open ocean),
or transported downwards to the deep ocean, where
carbon would be sequestered for long periods. One
of the main goals for future research must be to
elucidate which of the above carbon paths prevails in
this region.
Acknowledgements
This work has been supported through the
European Union (project CANIGO, number
MAS3-CT96-0060) and the Spanish government
(project COCA, number REN2000-U471-C02-02MAR). Part of this work was written while the
33
first author was at University of Wisconsin-Madison
with funding from the Secretarı́a de Estado de
Educación y Universidades of the Spanish Government. The satellite images used in this study were
received and processed at the Universidad de Las
Palmas de Gran Canaria. We are grateful to Susan
Lozier and Pierro Zanasca for providing the data to
produce Figs. 6c, d and 7. We also wish to thank
our reviewers for a number of useful comments and
suggestions, in particular one of them for his
exhaustive revision of the original manuscript that
greatly helped to improve it.
References
Agustı́, S., Duarte, C.M., Vaqué, D., Hein, M., Gasol, J.M., Vidal,
M., 2001. Food-web structure and elemental (C, N, and P)
fluxes in the eastern tropical North Atlantic. Deep-Sea Res., Part
II, Top. Stud. Oceanogr. 48, 2295 – 2321.
Allen, J.S., 1973. Upwelling and coastal jets in a continuously
stratified ocean. J. Phys. Oceanogr. 3, 245 – 257.
Arı́stegui, J., Montero, M.F., 2005. Temporal and spatial changes
in microplankton respiration and biomass in the Canary
Islands: the effect of mesoscale variability. J. Mar. Syst. 54,
65–82, this issue.
Arı́stegui, J., Hernández-León, S., Gómez, M., Medina, L., Ojeda,
A., Torres, S., 1989. Influence of the north trade winds on the
biomass and production of neritic plankton in Gran Canaria. Sci.
Mar. 53, 223 – 229.
Arı́stegui, J., Sangrà, P., Hernández-León, S., Cantón, M., Hernández-Guerra, A., Kerling, J.L., 1994. Island-induced eddies in the
Canary Islands. Deep-Sea Res. 41, 1509 – 1525.
Arı́stegui, J., Tett, P., Hernández-Guerra, A., Basterretxea, G.,
Montero, M.F., Wild, K., Sangrà, P., Hernández-León, S.,
Cantón, M., Garcı́a-Braun, J.A., Pacheco, M., Barton, E.D.,
1997. The influence of island-generated eddies on chlorophyll
distribution: a study of mesoscale variation around Gran
Canaria. Deep-Sea Res. 44, 71 – 96.
Arı́stegui, J., Hernández-León, S., Montero, M.F., Gómez, M.,
2001. The seasonal planktonic cycle in coastal waters of the
Canary Islands. Sci. Mar. 65 (S1), 51 – 58.
Badan-Dangon, A., Brink, K.H., Smith, R.L., 1986. On the
dynamical structure of the midshelf water column off northwest
Africa. Cont. Shelf Res. 5, 629 – 644.
Barton, E.D., 1987. Meanders, eddies and intrusions in the
thermocline front off northwest Africa. Oceanol. Acta 10,
267 – 283.
Barton, E.D., 1998. Eastern boundary of the North Atlantic:
Northwest Africa and Iberia. Coastal segment (18,E). In:
Robinson, A.R., Brink, K.H. (Eds.), The Sea, vol. 11. John
Wiley and Sons, New York, pp. 633 – 657.
Barton, E.D., Arı́stegui, J., Tett, P., Cantón, M., Garcı́a-Braun, J.,
Hernández-León, S., Nykjaer, L., Almeida, C., Almunia, J.,
34
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Ballesteros, S., Basterretxea, G., Escánez, J., Garcı́a-Weil, L.,
Hernández-Guerra, A., López-Laatzen, F., Molina, R., Montero,
M.F., Navarro-Pérez, E., Rodrı́guez, J.M., van Lenning, K.,
Vélez, H., Wild, K., 1998. The transition zone of the Canary
Current upwelling region. Prog. Oceanogr. 41, 455 – 504.
Barton, E.D., Basterretxea, G., Flament, P., Mitchelson-Jacob, E.G.,
Jones, B., Arı́stegui, J., Felix, H., 2000. Lee region of Gran
Canaria. J. Geophys. Res. 105, 17173 – 17193.
Barton, E.D., Flament, P., Dodds, H., Mitchelson-Jacob, E.G., 2001.
Mesoscale structures viewed by SAR and AVHRR near the
Canary Islands. Sci. Mar. 65 (S1), 167 – 175.
Basterretxea, G., 1994. Influencia de las estructuras oceanográficas
mesoescalares sobre la producción primaria en la región
Canaria. PhD thesis. Universidad de Las Palmas de Gran
Canaria, Spain. 113 pp.
Basterretxea, G., Arı́stegui, J., 2000. Mesoscale variability in
phytoplankton biomass distribution and photosynthetic parameters in the Canary-NW African coastal transition zone. Mar.
Ecol. Prog. Ser. 197, 27 – 40.
Basterretxea, G., Barton, E.D., Tett, P., Sangrà, P., Navarro-Pérez,
E., Arı́stegui, J., 2002. Eddy and deep chlorophyll maximum
response to wind-shear in the lee of Gran Canaria. Deep-Sea
Res. 49, 1087–1101.
Bindoff, N.L., McDougall, T.J., 1994. Diagnosing climate change
and ocean ventilation using hydrographic data. J. Phys. Oceanogr. 24, 1137 – 1152.
Braun, J.G., 1981. Estudios de Producción en aguas de las Islas
Canarias: II. Producción del Zooplancton. Bol. Inst. Esp.
Oceanogr. T6, 89 – 96.
Braun, J.G., Real, F., 1984. Estudios de producción en la bahı́a de
Antequera: una comparación con aguas oceánicas. Bol. Inst.
Esp. Oceanogr. 1, 134 – 140.
Brink, K.H., 1983. The near-surface dynamics of coastal upwelling.
Prog. Oceanogr. 12, 223 – 257.
Brink, K.H., Cowles, T.J., 1991. The coastal transition zone
experiment. J. Geophys. Res. 96, 14637 – 14647.
Broecker, W.S., Peng, T.H., 1982. Tracers in the Sea. Eldigio Press,
Palisades, New York. 690 pp.
Castro, C.G., Pérez, F.F., Holley, S., Rı́os, A.F., 1998. Characterization and modelling of water masses in the Northeast Atlantic.
Prog. Oceanogr. 41, 249 – 279.
Codispoti, R.C., Dugdale, R.C., Minas, H.J., 1982. A comparison of
the nutrient regimes off northwest Africa, Peru and Baja
California. Rapp. P-v. réun.-Comm. int. Explor. sci. mer.
Méditerr. 180, 184 – 201.
Csanady, G.T., 1977. Intermittent bfullQ upwelling in Lake Ontario.
J. Geophys. Res. 82, 397 – 419.
Dam, H.G., Zhang, X., Butler, M., Roman, M.R., 1995. Mesozooplankton grazing and metabolism at the equator in the
central Pacific: implications for carbon and nitrogen fluxes.
Deep-Sea Res. 42, 735 – 756.
Duarte, C.M., Agustı́, S., Arı́stegui, J., González, N., Anadón, R.,
2001. Evidence for a heterotrophic subtropical northeast
Atlantic. Limnol. Oceanogr. 46, 425 – 428.
Gabric, A.J., Garcı́a, L., van Camp, L., Nykjaer, L., Eifler, W.,
Schrimpf, W., 1993. Offshore export of shelf production in the
Cap Blanc giant filament as derived from CZCS imagery. J.
Geophys. Res. 98, 4697 – 4712.
Garcı́a-Muñoz, M., Arı́stegui, J., Pelegrı́, J.L., Antoranz, A., Ojeda,
A., Torres, M., 2005. Exchange of carbon and nutrients by an
upwelling filament off Cape Guir (NW Africa). J. Mar. Syst. 54,
83–95, this issue.
Gliwicks, Z.M., 1986. A lunar cycle in zooplankton. Ecology 67,
883 – 897.
González-Dávila, M., Santana-Casiano, J.M., 1999. In: Llinás,
O., Rodrı́guez de León, A., Siedler, G., Wefer, G. (Eds.),
ESTOC Data Report, Inf. Tech. Inst. Canario Cienc. Mar.,
vol. 7. Instituto Canario de Ciencı̀as Marinas, Telde.
González-Dávila, M., Santana-Casiano, J.M., Rueda, M.J., Llinás,
O., González-Dávila, E.F., 2003. Seasonal and interannual
variability of sea-surface carbon dioxide species at the European
station for time series in the ocean at the Canary Islands
(ESTOC) between 1996 and 2000. Glob. Biogeochem. Cycles
17, 1076.
González-Dávila, M., Santana-Casiano, M., De Armas, D., Escánez,
J., Guillén, P., 2005. The influence of island generated eddies on
the carbon dioxide system distribution south of the Canary
Islands. J. Mar. Syst., this issue.
Grall, J.R., Le Corre, P., Tréguer, P., 1982. Short-term variability of
primary production in coastal upwelling of Morocco. Rapp. P-v.
réun.-Comm. int. Explor. sci. mer. Méditerr. 180, 221 – 227.
Hagen, E., Zqlicke, C., Feistal, R., 1996. Near-surface structures
in the Cape Ghir filament off Morocco. Oceanol. Acta 19,
577 – 598.
Harvey, J., 1982. u-S relationships and water masses in the Eastern
North Atlantic. Deep-Sea Res. 29, 1021 – 1033.
Haynes, R., Barton, E.D., 1990. A poleward flow along the
Atlantic coast of the Iberian Peninsula. J. Geophys. Res. 95,
11425 – 11442.
Haynes, R., Barton, E.D., Pilling, I., 1993. Development,
persistence, and variability of upwelling filaments off the
Atlantic coast of the Iberia Peninsula. J. Geophys. Res. 98,
22681 – 22692.
Head, E.J.H., Harrison, W.G., Irwin, B.D., Horne, E.P.W., Li,
W.K.W., 1996. Plankton dynamics and carbon flux in an area
of upwelling off the coast of Morocco. Deep-Sea Res. 43,
1713 – 1738.
Hernández-Guerra, A., 1990. Estructuras oceanográficas observadas
en las aguas que rodean las Islas Canarias mediante escenas de
los sensores AVHRR y CZCS. PhD thesis, Universidad de Las
Palmas de Gran Canaria, Spain.
Hernández-Guerra, A., Nykjaer, L., 1997. Sea surface temperature
variability off northwest Africa: 1981–1989. Int. J. Remote
Sens. 18, 2539 – 2558.
Hernández-Guerra, A., Arı́stegui, J., Cantón, M., Nykjaer, L., 1993.
Phytoplankton pigments patterns in the Canary Islands as
determined using Coastal Zone Colour Scanner Data. Int. J.
Remote Sens. 14, 1431 – 1437.
Hernández-Guerra, A., López-Laatzen, F., Machı́n, F., de Armas,
D., Pelegrı́, J.L., 2001. Water masses, circulation and transport
in the eastern boundary current of the North Atlantic subtropical
gyre. Sci. Mar. 65 (S1), 177 – 186.
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Hernández-Guerra, A., Machı́n, F., Antoranz, A., Cisneros-Aguirre,
J., Gordo, C., Marrero-Dı́az, A., Martı́nez, A., Ratsimandresy,
A.W., Rodrı́guez-Santana, A., Sangrá, P., López-Laatzen, F.,
Parrilla, P., Pelegrı́, J.L., 2002. Temporal variability of mass
transport in the Canary Current. Deep-Sea Res., Part II, Top.
Stud. Oceanogr. 49, 3415 – 3426.
Hernández-León, S., 1988. Gradients of mesozooplankton biomass
and ETS activity in the wind-shear area as evidence of an island
mass effect in the Canary Island waters. J. Plankton Res. 10,
1141 – 1154.
Hernández-León, S., 1991. Accumulation of mesozooplankton in a
wake area as a causative mechanism of the bIsland-mass effectQ.
Mar. Biol. 109, 141 – 147.
Hernández-León, S., 1998. Annual cycle of epiplanktonic copepods
in Canary Island waters. Fish. Oceanogr. 7, 252 – 257.
Hernández-León, S., Llinás, O., Braun, J.G., 1984. Nota sobre la
variacion de la biomasa del mesozooplancton en aguas de
Canarias. Inv. Pesq. 48, 495 – 508.
Hernández-León, S., Arı́stegui, J., Gómez, M., Torres, S., Almeida,
C., Montero, M.F., Ojeda, A., 1998. Mesozooplankton metabolism and its effect on chlorophyll and primary production in
slope waters of the Canary Islands. Ann. Inst. Océanogr. Paris
74, 57 – 68.
Hernández-León, S., Postel, L., Arı́stegui, J., Gómez, M., Montero,
M.F., Torres, S., Almeida, C., Kqhner, E., Brenning, U., Hagen,
E., 1999. Large-scale and mesoscale distribution of plankton
biomass and metabolic activity in the Northeastern Central
Atlantic. J. Oceanogr. 55, 471 – 482.
Hernández-León, S., Almeida, C., Gómez, M., Torres, S., Montero,
I., Portillo-Hahnefeld, A., 2001a. Zooplankton biomass and
indices of feeding and metabolism in island-generated eddies. J.
Mar. Syst. 30, 51 – 66.
Hernández-León, S., Almeida, C., Yebra, L., Arı́stegui, J.,
Fernández de Puelles, M.L., Garcı́a-Braun, J., 2001b. Zooplankton abundance in subtropical waters: is there a lunar cycle?
Sci. Mar. 65, 59 – 64.
Hernández-León, S., Gómez, M., Pagazaurtundua, M., PortilloHahnefeld, A., Montero, I., Almeida, C., 2001c. Vertical
distribution of zooplankton in Canary Island waters: implications for export flux. Deep-Sea Res. 48, 1071 – 1092.
Hernández-León, S., Almeida, C., Portillo-Hahnefeld, A., Gómez,
M., Rodrı́guez, J.M., Arı́stegui, J., 2002. Zooplankton
biomass and indices of feeding and metabolism in relation
to a filament off the northwest African upwelling zone. J.
Mar. Res. 60, 327 – 346.
Huntsman, S.A., Barber, R.T., 1977. Primary production off
northwest Africa: the relationship to wind and nutrient
conditions. Deep-Sea Res. 24, 25 – 33.
INM, 1988. Mapa Eólico Nacional. Instituto Nacional de Meteorologı́a. 490 pp.
Isemer, H.J., Hase, L., 1987. The Bunker Climate Atlas of the North
Atlantic Ocean. Volume 2: Air–Sea Interactions. SpringerVerlag. 252 pp.
Jewell, P.W., 1994. Mass balance models of Ekman transport and
nutrient fluxes in coastal upwelling zones. Glob. Biogeochem.
Cycles 8, 165 – 177.
35
Jiménez, B., Sangrà, P., 1999. The influence of wind forcing on
Gran Canaria Island eddies generation. CANIGO Final Conference Book of Abstracts, Las Palmas de Gran Canaria, Spain,
pp. 82–83.
Joint, I., Inall, M., Torres, R., Figueiras, F.G., Alvarez-Salgado,
X.A., Rees, A.P., Woodward, E.M.S., 2001. Two lagrangian
experiments in the Iberian Upwelling System: tracking an
upwelling event and an off-shore filament. Prog. Oceanogr. 51,
221 – 248.
Kawase, M., Sarmiento, J.L., 1985. Nutrients in the Atlantic
thermocline. J. Geophys. Res. 90, 8961 – 8979.
Khripounoff, A., Vangriesheim, A., Crassous, P., 1998. Vertical and
temporal variations of particle fluxes in the deep tropical
Atlantic. Deep-Sea Res. 45, 193 – 216.
Knoll, M., Hernández-Guerra, A., Lenz, B., López-Laatzen, F.,
Machı́n, F., Mqller, T.J., Siedler, G., 2002. The eastern
boundary current system between the Canary Islands and the
African coast. Deep-Sea Res., Part II, Top. Stud. Oceanogr.
49, 3427 – 3440.
Kfrtsinger, A., Hedges, J.I., Quay, P.D., 2001. Redfield ratios
revisited: removing the biasing effect of anthropogenic CO2.
Limnol. Oceanogr. 46, 964 – 970.
Kostianoy, A.G., Zatsepin, A.G., 1996. The west African coastal
upwelling filaments and cross-frontal water exchange conditioned by them. J. Mar. Syst. 7, 349 – 359.
Laiz, I., Sangrà, P., Pelegrı́, J.L., Marrero-Dı́az, A., 2001. Sensitivity
of an idealised subtropical gyre to the eastern boundary
conditions. Sci. Mar. 65 (S1), 187 – 194.
Lee, S., Pelegrı́, J.L., Kroll, J., 2001. Slope control in western
boundary currents. J. Phys. Oceanogr. 31, 3349 – 3360.
Leetma, A., Bunker, A.F., 1978. Updated charts of the main
annual wind stress, convergences in the Ekman layers, and
Sverdrup transports in the North Atlantic. J. Mar. Res. 36,
311 – 322.
Lenz, J., 1982. Particulate organic matter off northwest Africa: size
distribution and composition in relation to shore distance and
depth. Rapp. P-v. réun.-Cons. int. expl. mer. Méditerr. 180,
234 – 238.
Lloyd, I.J., 1969. Primary production off the coast of northwest
Africa. J. Cons. int. Explor. Mer 33, 312 – 323.
Longhurst, A., 1998. Ecological Geography of the Sea. Academic
Press, San Diego, 398 pp.
Longhurst, A., Sathyendranath, S., Platt, T., Caverhill, C., 1995. An
estimate of global primary production in the ocean from satellite
radiometer data. J. Plankton Res. 17, 1245 – 1272.
Louanchi, F., Najjar, R.G., 2001. Annual cycles of nutrients and
oxygen in the upper layers of the North Atlantic Ocean. DeepSea Res., Part II, Top. Stud. Oceanogr. 48, 2155 – 2171.
Lozier, M.S., 1997. Evidence for large-scale eddy driven gyres in
the North Atlantic. Science 277 (5324), 361 – 364.
Lozier, M.S., Owens, W.B., Curry, R.G., 1995. The climatology of
the North Atlantic. Prog. Oceanogr. 36, 1 – 44.
Marrero, A., Pelegrı́, J.L., 1995. Water and nutrient transfer off
northwest Africa. Proceedings of the XXI International
Union of Geodesy and Geophysics, Boulder, Colorado,
p. B312.
36
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Marrero-Dı́az, A., Pelegrı́, J.L., Rodrı́guez-Santana, A., Sangrà, A.,
2001. Applicability of T–S algorithms to the Canary Islands
region. Sci. Mar. 65 (S1), 195 – 204.
McClain, C.R., Firestone, J., 1993. An investigation of Ekman
upwelling in the North Atlantic. J. Geophys. Res. 98,
12327 – 12339.
Minas, H.J., Codispoti, R.C., Dugdale, R.C., 1982. Nutrients and
primary production in the upwelling region off northwest
Africa. Rapp. P-v. réun.-Cons. int. explor. mer. Méditerr. 180,
148 – 183.
Mittelstaedt, E., 1991. The ocean boundary along the northwest
African coast: circulation and oceanographic properties at the
sea surface. Prog. Oceanogr. 26, 307 – 356.
Molina, R., Laatzen, F.L., 1989. Hidrografı́a en la región Canaria.
Campaña Canarias I. Bol. Inst. Esp. Oceanogr. 5, 71 – 86.
Molina, R., Cabanas, J.M., Laatzen, F.L., 1996. Corrientes e
hidrografı́a en la región Canaria: Campaña Canaria 9205. Bol.
Inst. Esp. Oceanogr. 12, 43 – 52.
Montero, M.F., 1993. Respiración y actividad ETS en microplancton marino. Variabilidad del ETS en aguas de Canarias.
PhD thesis. Universidad de Las Palmas de Gran Canaria,
Spain, 194 pp.
Navarro-Pérez, E., Barton, E.D., 1998. The physical structure of an
upwelling filament off the northwest African coast during
August 1993. S. Afr. J. Mar. Sci., vol. 19, pp. 61 – 74.
Navarro-Pérez, E., Barton, E.D., 2001. Seasonal and interannual
variability of the Canary Current. Sci. Mar. 65 (S1), 205 – 213.
Nykjaer, L., 1988. Remote sensing applied to the northwest African
upwelling are. Doctoral thesis, University of Copenhagen,
Denmark.
Nykjaer, L., Van Camp, L., 1994. Seasonal and interannual
variability of coastal upwelling along northwest Africa and
Portugal from 1981 to 1991. J. Geophys. Res. 99,
14197 – 14207.
Pacheco, M., Hernández-Guerra, A., 1999. Seasonal variability of
recurrent phytoplankton pigment patterns in the Canary Islands
area. Int. J. Remote Sens. 29, 1405 – 1418.
Palmén, E., Newton, C.W., 1969. Atmospheric circulation systems.
Academic Press, New York, 603 pp.
Patzert, W.C., 1969. Eddies in Hawaiin waters. Repport HIG-69-8,
Hawaii Institute of Geophysics, University of Hawaii at Manoa,
Honolulu, HI.
Pelegrı́, J.L., Richman, J.G., 1993. On the role of shear-mixing
during transient coastal upwelling. Cont. Shelf Res. 13,
1363 – 1400.
Pelegrı́, J.L., Sangrà, P., Hernández-Guerra, A., 1997. Heat gain
in the eastern North Atlantic subtropical gyre. In: Dı́az, J.I.
(Ed.), The Mathematics of Models for Climatology and
Environment, NATO ASI Series, vol. I 48. Springer-Verlag,
Berlin, pp. 419 – 436.
Pelegrı́, J.L., Marrero-Dı́az, A., Ratsimandresy, A.W., Antoranz, A.,
Cisneros-Aguirre, J., Gordo, C., Grisolı́a, D., HernándezGuerra, A., Láiz, I., Martı́nez, A., Parrilla, G., Pérez-Rodrı́guez,
P., Rodrı́guez-Santana, A., Sangrà, P., 2005. Hydrographic
cruises off northwest Africa: the Canary Current and the Cape
Ghir filament. J. Mar. Syst. 54, 39–63 (this issue).
Pérez, F.F., Mintrop, L., Llinás, O., González-Dávila, M., Castro,
C.G., Alvarez, M., Kfrtzinger, A., Santana-Casiano, M., Rueda,
M.J., Rı́os, A.F., 2001. Mixing analysis of nutrients, oxygen and
inorganic carbon in the Canary Islands region. J. Mar. Syst. 28,
183 – 201.
Pérez-Rodrı́guez, P., Pelegrı́, J.L., Marrero-Dı́az, A., 2001. Dynamical characteristics of the Cape Verde frontal zone. Sci. Mar. 65
(S1), 241 – 250.
Pinot, J.M., Jansá, J., 2001. Time variability of acoustic backscatter
from zooplankton in the Ibiza Channel (western Mediterranean).
Deep-Sea Res. 48, 1651 – 1670.
Pollard, R.T., Pu, S., 1985. Structure and circulation of the upper
Atlantic Ocean northeast of the Azores. Prog. Oceanogr. 14,
443 – 462.
Postel, L., 1990. Die Reaktion des Mesozooplanktons, speziell der
Biomasse, auf kqstennahen Auftrieb vor Westafrika. Doctoral
thesis, Institut fqr Meereskunde, Warnemqnde, Germany.
Ramp, S.R., Jessen, P.F., Brink, K.H., Niiler, P.P., Daggett, F.L.,
Best, J.S., 1991. The physical structure of cold filaments near
Point Arena, California, during June 1987. J. Geophys. Res. 96,
14859 – 14883.
Ratsimandresy, A.W., Pelegrı́, J.L., Marrero-Dı́az, A., HernándezGuerra, A., Antoranz, A., Martı́nez, A., 2001. Seasonal
variability of the upper warmwatersphere in the Canary Basin.
Sci. Mar. 65 (S1), 251 – 258.
Real, F., de Armas, J.D., Braun, J.G., 1981. Distribución espacial de
la clorofila a y del carbono orgánico particulado en aguas
costeras superficiales de las Islas Canarias. Bol. Inst. Esp.
Oceanogr. T6, 109 – 115.
Reed, R.K., 1977. On estimating insolation over the ocean. J. Phys.
Oceanogr. 7, 482 – 485.
Reid, J.L., 1994. On the total geostrophic circulation of the North
Atlantic Ocean: flow patterns, tracers, and transports. Prog.
Oceanogr. 33, 1 – 92.
Rodrı́guez, J.M., Hernández-León, S., Barton, E.D., 1999.
Mesoscale distribution of fish larvae in relation to an
upwelling filament off northwest Africa. Deep-Sea Res. 46,
1969 – 1984.
Rodrı́guez-Santana, A., Pelegrı́, J.L., Sangrà, P., 1999. Diapycnal
mixing in Gulf Stream meanders. J. Geophys. Res. 104,
25892 – 25912.
Rodrı́guez-Santana, A., Pelegrı́, J.L., Sangrà, P., Marrero-Dı́az,
A., 2001. Diapycnal mixing in a frontal system forced by an
oscillating deformation field. Phys. Chem. Earth (B) 26,
293 – 298.
Sangrà, P., 1995. Perturbatión de un flujo geofı́sico por un
obstáculo: aplicación a la isla de Gran Canaria. Doctoral thesis,
Universidad de Las Palmas de Gran Canaria, Spain.
Sangrà, P., Hernandez-Guerra, A., Marrero-Diaz, A., Martin, J.M.,
Martinez, A., Ratsimandresy, A.W., Rodriguez-Santana, A.,
2004. Life history of an anticyclonic eddy. J. Geophys. Res.
Santana-Casiano, J.M., González-Dávila, M., Laglera-Baquer,
L.M., Rodrı́guez-Somoza, M.J., 2001. Carbon dioxide system
in the Canary region during October 1995. Sci. Mar. 65, 41 – 49.
Schmitz, W.J., McCartney, M.S., 1993. On the North Atlantic
circulation. Rev. Geophys. 31, 29 – 49.
J.L. Pelegrı́ et al. / Journal of Marine Systems 54 (2005) 3–37
Schmitt, R.W., Bogden, P.S., Dorman, C.E., 1989. Evaporation
minus precipitation and density fluxes for the North Atlantic. J.
Phys. Oceanogr. 19, 1208 – 1221.
Schulz, S., 1982. A comparison of primary production in upwelling
regions off northwest and southwest Africa. Rapp. P-v. réun.Cons. int. explor. mer. Méditerr. 180, 202 – 204.
Siedler, G., Onken, R., 1996. Eastern recirculation. In: Krauss, W.
(Ed.), The Warmwatersphere of the North Atlantic Ocean.
Gebrqder Borntraeger, Berlin, pp. 339 – 364.
Siedler, G., Khuhl, A., Zenk, W., 1987. The Madeira mode water. J.
Phys. Oceanogr. 17, 1561 – 1570.
Smith, S.L., Whitledge, T.E., 1982. Regeneration of nutrients by
zooplankton and fish off northwest Africa. Rapp. P-v. réun.Cons. int. explor. mer. Méditerr. 180, 206 – 208.
Speth, P., Detlefsen, H., 1982. Meteorological influences on
upwelling off northwest Africa. Rapp. P-v. réun.-Cons. int.
explor. mer. Méditerr. 180, 29 – 34.
Stommel, H., 1979. Determination of water mass properties of water
pumped down from the Ekman layer to the geostrophic flow
below. Proc. Natl. Acad. Sci. U. S. A. 76, 3051 – 3055.
Stramma, L., 1984. Geostrophic transport in the Warm Water
Sphere of the eastern subtropical North Atlantic. J. Mar. Res. 42,
537 – 558.
Stramma, L., Siedler, G., 1988. Seasonal changes in the North
Atlantic subtropical gyre. J. Geophys. Res. 93, 8111 – 8118.
Strub, P.T., Kosro, P.M., Huyer, A., 1991. The nature of cold
filaments in the California Current system. J. Geophys. Res. 96,
14743 – 14768.
Takahashi, T., Wanninkhof, R.H., Feely, R.A., Weiss, R.F.,
Chipman, D.W., Bates, N., Olafsson, J., Sabine, C., Sutherland, S.C., 1999. Net sea–air CO2 flux over the global
37
oceans: An improved estimate based on the air sea–air pCO2
difference. Proceedings of the 2nd International Symposium
CO2 in the Oceans, Tsukuba, January 1999, CGER-1037-’99,
CGER/NIES.
Thiriot, A., 1978. Zooplankton communities in the west African
upwellling area. In: Boje, R., Tomczak, M. (Eds.), Upwelling
Ecosystems. Springer-Verlag, Berlin, pp. 32 – 61.
Tomczak, M., 1981. An analysis of mixing in the frontal zone of
South and North Atlantic central water off northwest Africa.
Prog. Oceanogr. 10, 173 – 192.
Van Camp, L., Nykjaer, L., Mittelstaed, E., Schlittenhard, P., 1991.
Upwelling and boundary circulation off northwest Africa as
depicted by infrared and visible satellite observations. Prog.
Oceanogr. 26, 357 – 402.
Van Lenning, K., 2000. Variability in biomass and structure
of phytoplankton populations in the Canary Islands
waters, as determined by HPLC analyses of pigments.
PhD thesis. Universidad de Las Palmas de Gran Canaria,
Spain, 270 pp.
Vives, F., 1975. Zooplancton et Hydrographie des environs du Cap
Blanc (NO. de l’Afrique). C.M.-Int. Counc. Explor. Sea 1975/
L:10.6.
Wanninkhof, R., 1992. Relationship between wind speed and gas
exchange over the ocean. J. Geophys. Res. 97, 7373 – 7382.
Wooster, W.S., Bakun, A., McLain, D.R., 1976. The seasonal
upwelling cycle along the eastern boundary of the North
Atlantic. J. Mar. Res. 34, 131 – 141.
Worthington, L.V., 1959. The 18 8C water in the Sargasso Sea.
Deep-Sea Res. 5, 297 – 305.
Zenk, W., Klein, B., Schrfder, M., 1991. Cape Verde frontal zone.
Deep-Sea Res. 38, S505 – S530.