GLOBAL BIOGEOCHEMICAL CYCLES, VOL. 17, NO. 3, 1076, doi:10.1029/2002GB001993, 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
Melchor González-Dávila and J. Magdalena Santana-Casiano
Chemistry Department, Faculty of Marine Science, University of Las Palmas de Gran Canaria, Las Palmas de Gran Canaria,
Spain
Maria-José Rueda and Octavio Llinás1
Instituto Canario de Ciencias Marinas, Gobierno de Canarias, Las Palmas de Gran Canaria, Spain
Enrique-Francisco González-Dávila
Statistical Department, Faculty of Mathematics, University of La Laguna, Tenerife, Spain
Received 27 September 2002; revised 21 April 2003; accepted 19 May 2003; published 15 July 2003.
[1] Seasonal patterns in hydrography, partial pressure of CO2, fCO2, pHt, total alkalinity,
AT, total dissolved inorganic carbon, CT, nutrients, and chlorophyll a were measured in
surface waters on monthly cruises at the European Station for Time Series in the Ocean at
the Canary Islands (ESTOC) located in the northeast Atlantic subtropical gyre. With
over 5 years of oceanographic data starting in 1996, seasonal and interannual trends of
CO2 species and air-sea exchange of CO2 were determined. Net CO2 fluxes show this
area acts as a minor source of CO2, with an average outgassing value of 179 mmol CO2
m2 yr1 controlled by the dominant trade winds blowing from May to August. The
effect of short-term wind variability on the CO2 flux has been addressed by increasing airsea fluxes by 63% for 6-hourly sampling frequency. The processes governing the monthly
variations of CT have been determined. From March to October, when CT decreases,
mixing at the base of the mixed layer (11.5 ± 1.5 mmol m3) is compensated by air-sea
exchange, and a net organic production of 25.5 ± 5.7 mmol m3 is estimated. On an
annual scale, biological drawdown accounts for the decrease in inorganic carbon from
March to October, while mixing processes control the CT increase from October to the
end of autumn. After removing seasonality variability, f CO2sw increases at a rate
of 0.71 ± 5.1 matm yr1, and as a response to the atmospheric trend, inorganic carbon
INDEX TERMS: 1615 Global Change:
increases at a rate of 0.39 ± 1.6 mmol kg1 yr1.
Biogeochemical processes (4805); 4806 Oceanography: Biological and Chemical: Carbon cycling; 4227
Oceanography: General: Diurnal, seasonal, and annual cycles; 9325 Information Related to Geographic
Region: Atlantic Ocean; KEYWORDS: Atlantic Ocean, carbon dioxide, time series
Citation: González-Dávila, M., J. M. Santana-Casiano, M.-J. Rueda, O. Llinás, and E.-F. González-Dávila, 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, Global Biogeochem. Cycles, 17(3), 1076, doi:10.1029/2002GB001993, 2003.
1. Introduction
[2] The global climate system is significantly influenced
by the ocean. The capacity of the oceans for uptake of CO2
depends on inorganic carbon chemistry and also to a great
extent on many different factors such as hydrography,
circulation of water masses, mixed layer dynamics, wind
stresses and biological processes in the ocean [Broecker and
Peng, 1982]. The ocean is estimated to have taken up about
1
Also at Physical Department, Faculty of Marine Science, University of
Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain.
Copyright 2003 by the American Geophysical Union.
0886-6236/03/2002GB001993$12.00
30% of the CO2 released into the atmosphere through
anthropogenic activities during the last few decades. The
North Atlantic Ocean, with high-latitude oceanic regions of
deep water formation, midlatitude sites of mode water
formation, and a subtropical oligotrophic ocean, is thought
to be a large sink for atmospheric CO2 [Tans et al., 1990;
Takahashi et al., 1993, 2002]. The Canary Islands area is a
transitional zone between the northwest African coastal
upwelling region and the open ocean oligotrophic waters
of the subtropical gyre. They present a barrier to the
relatively weak equatorward flow of the Canary Currents
and to the flow of trade winds, giving rise to a variety of
mesoscale phenomena south of the islands which have
strong implications on the biogeochemical cycles of the
2-1
2-2
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
region [González-Dávila et al., 2003]. As a result of the
seasonal variability of the Trade winds, upwelling along the
northwest African coast north of 25N is strongest during
the summer months [Van Camp et al., 1991], and filaments
of upwelled water have been observed to advect several
hundred kilometers westward. These filaments can be
observed from the Coastal Zone Colour Scanner (CZCS)
observations in the region [Davenport et al., 1999]. However, upwelling filaments normally reach the islands in the
south, and European Station for Time Series in the Ocean at
the Canary Islands (ESTOC) normally does not receive
filaments because of the barrier function of the eastern
islands [Davenport et al., 1999, 2002]. In this study we
present the results of 5 years’ worth of investigation into the
carbon dioxide system at ESTOC positioned at 29100N,
15300W, about 100 km north of the island of Gran Canaria.
ESTOC was initiated in 1994 as a joint Spanish-German
initiative. Samples are taken monthly at the station and are
supplemented by current meter and sediment trap mooring
at the site. The monthly sampling is carried out by regional
cruises using the Canary Island governmental oceanographic
vessel ‘‘Taliarte,’’ and, several times a year, additional
German vessels in the area are used in addition. The aim of
these German cruises is the validation of time-series observations over a wider region. ESTOC was also a reference
station for the EU project CANIGO (1996– 1999).
[3] One of the main objectives of the station is to
investigate biogeochemical cycles in this region in order
to provide insight into the processes controlling the flux of
carbon and associated elements in the ocean on seasonal
and interannual timescales. From October 1995, monthly
determination of oxygen, nutrients, and chlorophyll concentration by the ICCM (Instituto Canario de Ciencias
Marinas) and the fugacity of CO2 in the atmosphere and
surface seawater, together with pH and total alkalinity in the
water column by the Marine Chemistry group (QUIMA) of
ULPGC were determined in order to elucidate the behavior
of carbon dioxide in the area. In this first paper considering
monthly data for the ESTOC site, we present inorganic
carbon distributions from 1996 to 2000 together with
quantification of the role the various processes play in the
seasonal evolution of carbon parameters at the site. The
special role played by mixing and biological activity will be
addressed. A special effort was made clearly to define longterm trends for the different parameters affecting the carbon
dioxide system. We compared the results obtained with
other Time Series stations established, such as the Bermuda
Atlantic Time-series Study (BATS), located in the northwest
Sargasso Sea and in the return flow of the subtropical gyre
system [Michaels and Knap, 1996], and the Hawaii Ocean
Time series program (HOT) located at the North Pacific
Subtropical Gyre [Winn et al., 1994; Karl et al., 2001].
More information on the goals of ESTOC and the organizational structure can be found online www.iccm.rcanarias.
eswww.Ifm.uni-kiel and www.allgeo.uni-bremen.de).
2. Methods and Materials
[4] Hydrographic and biogeochemical data were collected
at the ESTOC station each month from October 1995
onwards on board the following research vessels: BO
Taliarte (36 cruises), FS Poseidon (18 cruises), FS Meteor
(9 cruises), FS Victor Hensen (9 cruises), and BO Navarro
(1 cruise). Hydrographic properties were measured with a
self-contained CTD-SIS to a maximum depth of 1000 m
(ICCM) and an OTS 1500 CTD probe (UOL) down to
3000 m for Victor Hensen cruises; 5- and 10-LNiskin bottles
were used to collect water for samples. Throughout Meteor
and Poseidon cruises, a MKIIIBNeil Brown CTD was used,
mounted in a General Oceanic rosette sampler to which
21 10 L Niskin bottles were attached. Comments on the
procedures can be found in the ESTOC Data Reports
[Llinás et al., 1997, 1999].
2.1. Nutrients
[5] Samples for nutrients were analyzed in the lab following the World Ocean Circulation Experiment (WOCE)
standards [WOCE, 1994]. Nutrients were determined by
colorimetric methods, using a Technicon AutoAnalyzer
AAII. For silicates, a modified Hansen and Grasshoff
[1983] method was used, in which ß-silicomolybdenic acid
is reduced with ascorbic acid. Nitrate was determined after
reduction to nitrite in a Cd-Cu column. The standard
deviation for duplicates was 0.07 mmol/L for silicate, 0.06
mmol/L for nitrate, and 0.01 mmol/L for phosphate. This is
equivalent to 0.3%, 0.5%, and 0.8% (full-scale) reproducibility respectively.
2.2. Chlorophyll a
[6] Chlorophyll a was sampled from 200 m to the surface
using 1-L polypropylene bottles. Phytoplankton pigments
were collected by filtration of 500 mL of the water sample
using Whatman GF/F 47-mm glass microfiber filters
applying the filtration pressure recommended by the
JGOFS-WOCE protocols [WOCE, 1994]. Each filter was
saved in 10-mL glass tubes and frozen subsequently at
20C. The day before analyzing, 10-mL acetone (90%)
was added to each filter to extract the pigments. The filters
were left in the fridge for 24 hours and then chlorophyll a
was measured using fluorometric analysis according to the
methodology described by Welschmeyer [1994], using a
fluorometer TURNER2 10-AU-000.
2.3. The pH
[7] From October 1995 to June 1997, the pHT in total
scale (mol (kg sw)1) was measured at 25C using the
potentiometric technique. The electrodes used to measure
the emf of the sample consisted of a ROSS2 glass pH
electrode and an Orion double junction Ag, AgCl reference
electrode, connected to an Orion2 720A pH meter. The
electrodes were calibrated by using tris/HCl buffer in
synthetic seawater with salinity 35. The pHT of the unknown seawater samples was determined according to
standard operation procedures (SOP 6 [Dickson and Goyet,
1994]). Seawater measurements of different CRM (certified
reference material for CO2 determination) samples (n = 54,
batch 34) repeatedly gave a standard deviation of ±0.003
units.
[8] After June 1997, the pHT was measured following the
spectrophotometric technique of Clayton and Byrne [1993]
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
2-3
Table 1. Relationship of Differences Related to Carbon Dioxide Properties Observed Between ESTOC and
BATS Sites
Property
Position
SST range
SSS range
Maximum mixed layer depth
Average NAT, mmol kg1
DNCT annual cycle, mmol kg1
Df CO2 annual cycle, matm
PP, mol C m2 yr1b
Export POC, mol C m2 yr1b
NCT increase, mmol kg1 yr1
f CO2 increase, matm yr1
BATSa
ESTOC
0
0
2 910 N, 1530 W
18 – 24C
36.6 – 36.9
150 m
2291.5 ± 2.6
25 (2007 – 1981)
60 – 80
14.5
1.1
0.39 ± 1.6 (1996 – 2000)
0.71 ± 5.1
0
3150 N, 64100W
20 – 29C
36.2 – 36.9
250 m
2284.0 ± 4.2
45 (2065 – 2010)
90 – 100
11.9
0.2
1.6 ± 5.8 (1988 – 1998)
1.4 ± 10.7
a
Bates [2001].
Neuer et al. [2002].
b
using the m-cresol purple indicator [Dickson and Goyet,
1994]. A system similar to that described by Bellerby et al.
[1995] was developed in our lab. The pHT measurements
were carried out using a Hewlett Packard Diode Array
spectrophotometer in a 25C-thermostatted 1-cm flow-cell
using a Peltier system. A stopped-flow protocol was used to
analyze seawater previously thermostatted to 25C for a
blank determination at 730 nm, 578 nm, and 434 nm. The
flow was restarted, and the indicator injection valve
switched on to inject 10 mL dye through a mixing coil
(2 m). Three photometric measurements were carried out
for each injection after mixing along the coil in order to
remove all dye effect from the seawater pHT reading.
Repeatedly, seawater measurements of different CRM
samples (batches 35 and 42, provided by A. Dickson,
Scripps Institution of Oceanography) give a standard
deviation of ±0.0015 (n = 59). From April 1999 to
September 1999, carbon dioxide variables were not
determined at the ESTOC due to logistical problems.
2.4. Total Alkalinity
[9] The total alkalinity of seawater (AT) was determined
by titration with HCl to carbonic acid end point using two
similar potentiometric systems, described in greater detail
by Mintrop et al. [2000]. The total alkalinity of seawater
was evaluated from the proton balance at the alkalinity
equivalence point, pHequiv = 4.5, according to the exact
definition of total alkalinity [Dickson, 1981]. The performance of the titration systems was monitored by titrating
different samples of certified reference material (CRM,
batches 32, 35, and 42) which have known inorganic carbon
and AT values. The concurrence between our data and
certified values was within ±1.5 mmol kg1.
2.5. Total Dissolved Inorganic Carbon
[10] Total inorganic carbon (CT) was computed from
experimental values of pHT and total alkalinity, using the
carbonic acid dissociation constants of Mehrbach according
to Dickson and Millero [1987]. This set of constants
presented the closest matching of CT(pH, AT) calculations
and certified CT values for CRM, batch 42. On average, the
CT(pH, AT) values are 3 mmol kg1 higher than certified
values with a CT residual of ±3 mmol kg1 (n = 52). CT
data were corrected accordingly. Moreover, the readjusted
Mehrbach dissociation constants best represent the fugacity
of carbon dioxide determined from the combination of AT
and pHT in surface seawater over the range of sea-surface
temperature and salinities.
2.6. Experimental f CO2
[11] After January 1997, fugacity of carbon dioxide
( fCO2) in air and in surface seawater was determined using
a flow system similar to the one designed by Wanninkhof
and Thoning [1993] and developed by Frank J. Millero’s
group at the University of Miami. The equilibrator used was
based on the design by R. F. Weiss and described by Butler
et al. [1988]. The concentration of CO2 in the air and in the
equilibrated air sample was measured with a differential,
nondispersive, infrared gas analyzer supplied by LI-COR2
(LI-6262 CO2/H2O Analyzer). The samples were measured
wet and the signal corrected for water vapor using the water
channel of the LI-COR detector. The instrument was
operated in the absolute mode and gathered CO2 concentrations directly from the instrument. Atmospheric air is
pumped at the bow of the ship and measured every hour.
The system was calibrated by measuring two different
standard gases with mixing ratios of 348.55 and 520.83 ppm
CO2 in the air. The National Oceanographic Organization
provided these calibrated standards and Atmospheric
Administration and they are traceable to the World
Meteorology Organization scale. Our system has a precision
of up to 1 matm and is accurate, for standard gases, to
2 matm. Fugacity of CO2 in seawater was calculated from
the measured xCO2.
3. Results and Conclusions
3.1. Hydrographic Variability
[12] The ESTOC site undergoes a seasonal variation in
surface temperature of 4– 6C (from 18C to 24C), which
is 4 – 5C less than that reported for BATS of 8– 10C
[Michaels et al.,1994; Bates, 2001] (Table 1) and is in a
similar range (from 23C to 27C) as for HOT [Karl et al.,
2001]. The ESTOC presents the lowest temperature of these
three midlatitude stations. The combined effect of the
Canary Current and the northwest African coastal upwelling
is responsible for this situation, as can be observed from
surface thermal images and distributions (Figures 1 and 2).
2-4
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 1. Sea surface temperature from the Advanced
Very High Resolution Radiometer (AVHRR) sensor of the
National Oceanic and Atmospheric Administration (NOAA)
for the Canary Islands area, showing the position of the
ESTOC site for November 10, 1999. See color version of
this figure at back of this issue.
[13] At ESTOC, a seasonal cycle related to surface
cooling and wintertime convective mixing increases the
depth of the mixing layer. During winter, sea surface
temperature (SST) varies between 17.9C (February
1999) and 19.5C, with the lowest values observed during
the winter of 1998 (Figures 2 and 3). The seasonal
thermocline is located at around 35– 150 m (following
Brainerd and Greg [1995]) reaching its maximum stratification in September – October, when the SST reaches
24.1C (in 1997) and only 22.7C (in 1996). The fluctuation
of the mixed-layer depth at the ESTOC ranged from 150 m
(year 2000) and 100 m from February to April, to <30 m
during summer. Again, winter mixing at BATS reached
much lower depths than at ESTOC (>250 m, Table 1) and
was also favored by stronger wind in this area during this
season. During the first 3 years, 1996– 1998, the mixed
layer was related to the 19– 20C isotherms, within a range
of 40– 100 m. During the 1999 – 2000 period, the range
became longer and cooler, partly because of the18 – 19C
isotherm (Figure 2). In the subsurface seawater, from 1996
to 1998, a relatively stable thermal distribution was
observed, whereas during the 1999 – 2000 period, a larger
oscillation was seen. The 18C isotherm reached a depth of
less than 25 m in February 1999 and over 150 m in January
2000.
[14] The related mixed layer cycle gives rise to a nutrient
pump over the euphotic zone, favoring phytoplankton
bloom and a period of increased primary production in the
mixing layer. Winter mixing increases the nitrate and phosphate depleted water over the mixed layer in the first 75 m to
1.5 mmol kg1 and 0.11 mmol kg1, respectively, whereas
during the rest of the year the levels were close to detection
limit (Figure 1). At 200 m, the typical value of nitrate +
nitrite (N + N) is 5 mmol L1, with maximum and minimum
values related to the displacement of the coupled nitracline
and thermocline. However, silicate distribution in the surface waters shows high values (around 2 mmol L1) where
N + N and P are very low; thus, this nonmetabolic supply is
related to the arrival of Saharan dust at ESTOC [PérezMarrero, 1998]. This nutrient supply during winter mixing
increases chlorophyll a concentration to 0.25 mg L1 over
the whole water column. In spring, bloom is reduced on the
surface, giving the deep of chlorophyll maximum (DCM) at
around 100 m with values of 0.3 mg L1, decreasing in
depth and in value through autumn to 75 m and 0.2 mg L1,
respectively. As a rule, the highest chlorophyll a concentrations are associated with the 18C and 20C isotherms.
When these isotherms reach the sea surface, the highest
chlorophyll a values are also to be found in surface water.
The highest chlorophyll a values (0.7 mg C L1) were
observed during 1999, when the 18C isotherm reaches the
surface water. The year 1998 gave different readings, with
the highest values being found during summer-autumn. In
addition, during 1999, a second chlorophyll sub-maximum
was observed over the same period. This situation may be
due to remnants of upwelling filaments but can also be due
to local (or mesoscale) enhancements of production
(breaking internal wave, passage of eddy or meddy)
(following Davenport et al. [1999, 2002]) (Figure 1).
Moreover, the DCM is associated with the nutricline and the
maximum of surface oxygen (data not shown), which
follows a similar seasonal cycle to that observed in BATS
[Michaels et al., 1994; Michaels and Knap, 1996]).
3.2. Average Monthly Evolution of Surface pHT, AT,
and CT (Mmol kg1) for the ESTOC Station
[15] The seasonal variability of surface pHT (Figure 3) is
quite large, from 8.005 in February – March, up to 8.05 in
September – October. This variability is in the range found
(8.00– 8.05) for the Canary Azores and Gibraltar (CANIGO)
area (A. F. Rios et al., Seasonal sea-surface carbon dioxide
in Azores area, submitted to Deep Sea Research, 2002)
(hereinafter referred to as Rios et al., submitted manuscript,
2002).
[16] Surface alkalinity (Figure 3) shows an average variation of 25 mmol kg1, quite similar to levels found for the
BATS site of 30 mmol kg1 [Bates, 2001]. The lowest
values were recorded in March 1996 (2398.6 mmol kg1)
and the highest in September 1997 (2424.2 mmol Kg1),
coinciding with the data for salinity distribution. The values
of normalized AT to a constant salinity of 35 (NAT), which
removes the effects of evaporation and precipitation, are
2285.0 mmol kg1 (March 2000) and 2296.4 mmol kg1
(September 1997). A mean NAT value of 2291.5 ± 2.6 mmol
kg1 is found at the ESTOC, which coincides with the
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
2-5
Figure 2. Distribution of (a) temperature and chlorophyll a concentration (mg m3) for the first 200 m
on ESTOC from 1996 to 2000 and (b) salinity and nitrate + nitrite (N + N) (mmol kg1) for two
representative years, 1996 and 1997. See color version of this figure at back of this issue.
average value obtained for the North Atlantic, 2291 mmol
kg1 [Millero et al., 1998] and is close to that obtained for
BATS, 2284.0 ± 4.2 mmol kg1 [Bates, 2001, Table 1].
[17] Total inorganic carbon in the surface layer at ESTOC,
CT, (Figure 3) varied seasonally between 2110 mmol kg1
in May and 2080 mmol kg1 in September –October when
the lowest CT concentrations were obtained. CT subsequently
increased again, reaching concentrations close to the annual
average concentration in December. The mean value was
2097.4 ± 5.6 mmol kg1. The values of normalized CT to a
constant salinity of 35 (NCT) were 2006.7 mmol kg1 (March
1997) and 1980.7 mmol kg1 (September 1997). The mean
value of NCT is 1992.4 ± 6.3. CT seasonal variation was only
20 –25 mmol kg1yr1, similar to the range found in the
Azores area, where values of 19 mmol kg1yr1 have been
recorded (Rios et al., submitted manuscript, 2002) and other
oligotrophic regions such as in the North Pacific Subtropical
gyre where a variation of 15 mmol kg1yr1 has been
observed [Winn et al., 1994]. However, it was lower than
the variation of 45 mmol kg1yr1 found in BATS [Bates et
al., 1996]. The values of normalized CT found in BATS
during the period between 1988 – 1993 were higher than the
ESTOC readings, ranging from 2010 mmol kg1 to
2065 mmol kg1.
[18] A characteristic seasonal pattern of CT distribution
was observed each year similar to that observed at BATS and
HOT, with a spring (April – May) maximum of CT in superficial waters, a spring-summer (May – August) decrease of
CT, a late summer-autumn (September– October) minimum,
and an increase of CT during autumn-winter (November–
2-6
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
March). The annual variability of CT at the ESTOC site is
strongly correlated with mixed layer temperature. CT in the
upper ocean is primarily controlled by the factors which
govern salinity and other nonconservative processes, such as
photosynthesis, oxidation of organic matter, air-sea exchange
of carbon dioxide, and precipitation and dissolution of
calcium carbonate [Lee et al., 2000]. Surface CT algorithms
are useful for studying several aspects of the marine carbon
system. Empirical algorithms of CT with SST have been
presented by Lee et al. [2000] for the major ocean basins
using various available data sets. We have applied an
algorithm to the 5 years’ worth of data at ESTOC. Corrections
were applied to account for surface water CT increases due to
rising atmospheric CO2 concentration (see below). A single
relationship was determined over the period of a year
between total normalized inorganic carbon at a constant
salinity of 35, NCT and SST, for surface data at ESTOC,
where the SST changed between 18 and 24C.
NCT ¼ 1972ð1Þ 4:23ð0:61ÞðT 25Þ
þ 0:102ð0:075ÞðT 25Þ2
r2 ¼ 0:97; residual 2:3 mmol kg1 N ¼ 55:
ð1Þ
In this area, part of the seasonal surface CT variation is
removed by normalizing the results to a constant salinity.
NCT increases during seasonal cooling, due to the convective
mixing of deep waters rich in NCT. In the springtime, surface
warming facilitates phytoplankton bloom, decreasing surface CT. However, this contribution to the oligotrophic
station is very small (nitrate + nitrite concentration never
reaches 0.5 mmol kg1 in the first 50 m), and thus changes in
NCT are inversely correlated with the SST. In order to
compare our data with those at BATS presented by Lee et al.
[2000], a second relationship with a reference temperature of
29C was used.
NCT ¼ 1957ð5Þ 3:42ð1:30ÞðT 29Þ
þ 0:102ð0:075ÞðT 29Þ2 :
ð2Þ
[19] Our values at temperatures higher than 29C are
15 mmol kg1 higher than the reading of 1940 mmol kg1
determined with the BATS data. Because the various factors
contributing to the surface CT concentration associated with
the SST (transport, biological, and thermodynamic effects)
differ geographically, these variations can be attributed to
the west and east position in the North Atlantic subtropical
gyre, with relatively cold surface waters, as was hypothesized by Lee et al. [2000].
3.3. Average Monthly Evolution of fCO2sw , f CO2air,
and FCO2 (mmol m2d1) for the ESTOC Station
[20] A strong seasonal variability was found for f CO2sw
(Figure 3) which fluctuated by 60– 80 matm. The f CO2sw
2-7
annual cycle presents the characteristic minimum values
(320– 330 matm) in winter and maximum in summer (390–
400 matm). The interannual increase of fCO2sw is 1.3 ±
20 matm yr1 close to the value reported for BATS of
1.4 matm yr1 [Bates et al., 1996]. The fCO2air is slightly
higher in winter than in summer, with a variation of 7 matm
and an interannual increase of 1.5 matm yr1. When we
compared the ESTOC fCO2sw data with the BATS fCO2sw
data, the seasonal pattern was the same in the two stations.
BATS presented a higher seasonal variability for pCO2sw of
90– 100 matm during the annual cycle. The highest values of
pCO2 were found in late summer (390 – 410 matm) and the
lowest values (310 – 320 matm) in winter.
[21] Variability in seawater fCO2 is closely linked to
temperature variability, reflecting a dominant control of
temperature over the f CO2 in the Canary Island region (Rios
et al., submitted manuscript 2002). Following Abraham and
Ledolter [1983], time series of fCO2 in seawater and SST
for ESTOC were mathematically related by decomposing
the series by trend, seasonal component, and error using
harmonic functions,
f CO2 ¼ 161:745 þ 0:1024 * t þ 9:4247 SST 8:064 sinðpt=6Þ
3:448 cosðpt=6Þ 2:993 sinðpt=3Þ
1:578 cosðpt=3Þ þ 1:321 sinð2pt=3Þ;
ð3Þ
with r2 = 0.979 and n = 60. Here t is time in months after
January 1996. The linear trend of f CO2, between 2
consecutive years at a constant temperature, gives an
average annual increase of 1.23 matm, with a 95% degree
of confidence,
[22] The surface water in the ESTOC station is CO2
undersaturated from December to May, acting as a CO2
sink and oversaturated from June to November, acting as a
source of CO2 [González-Dávila et al., 1998; SantanaCasiano et al., 2001]. Flux of CO2 is primarily driven by
wind strength and the difference in the CO2 fugacity in the
surface mixed layer and that of the overlying atmosphere.
The Canary Islands wind data (Figure 4) are affected by the
permanent Azores anticyclone, which provides predominant
trade winds blowing northeast between the end of May and
the beginning of September. During the rest of the year, the
wind is quite constant with values around 5 ± 2 m s1.
However, it is during the presence of the predominant Trade
Winds when the speed increases reach 12 ± 3 m s1, just
when the surface water is becoming CO2 oversaturated.
These combined effects make, on annual scales, the ESTOC
station act as a minor source of CO2 with a rate of 0.3 ±
0.4 mmol m2 d1 (110 mmol m2 yr1) computed with
daily averaged wind speed, considering only the sampling
data and following Wanninkhoff [1992]. Wanninkhof and
McGillis [1999] k-wind speed relationship gives us a source
of 140 mmol m2 yr1. The maximum outgassing rate is in
Figure 3. (opposite) Time series of surface properties at the ESTOC site from 1996 to 2000. (a) Temperature (C).
(b) Salinity. (c) The pHT in total scale and 25C. (d) AT (solid circles) and normalized to a constant salinity of 35 AT (NAT,
open circles) in mmol kg1. (e) CT (solid circles) and normalized to a constant salinity of 35 CT (NCT, open circles) in mmol
kg1. (f) Atmospheric f CO2 (open circles) and seawater f CO2 (solid circles) in matm.
2-8
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 4. (a) Wind (m s1) data determined at a station located at the International Airport of Gran
Canaria every 6 hours. Open circles are the data determined on board for each CO2 sampling time. (b)
Flux of CO2 determined considering both wind data each 6 hours and f CO2 data after equation (3), and
considering only CO2 experimental data and wind data as in Figure 4a (open circles).
August (8 – 10 mmol m2 d1) and the maximum rate of
ingassing is in February –March (6 mmol m2 d1).
Moreover, the high degree of wind speed variability
observed in Figure 4 for this area will affect the final airsea CO2 flux due to the nonlinear relationship between wind
speed and the CO2 gas transfer coefficient. In order to gain
insight into the wind variability effect [Bates and Merlivat,
2001] on the carbon dioxide flux for the ESTOC site, the
following approach was applied. The SST computed each
6 hours from AVHRR images at ESTOC site was provided
by the Laboratoire d’Etudes en Gèophysique et Océanographie Spatiale (LEGOS, Toulouse), while salinity values
were linearly interpolated to the time over which the wind
data were sampled. Applying the relationship between
f CO2,sw, temperature and time (equation (3)), the f CO2,sw
was computed at each wind time. The f CO2,air was linearly
interpolated between known values. Following the Wanninkhoff [1992] relationship, the estimated flux was computed and presented together with the flux on the sampling
days at the ESTOC in Figure 4. Short time variability from
days to weeks was expected due to warming and cooling of
the surface layer associated to the diurnal thermal cycle.
These effects were not considered in the previous study
where we looked at sampling days on a monthly frequency.
Following Bates et al. [1998] results on CO2 fluxes near
Bermuda, the high level of daily variation in CO2, 10– 14
mmol m2, is related to daily wind variability as high as
10 m s1. Moreover, in a daily cycle, fCO2 variability of
around 5 – 7 matm was strongly correlated to the temperature increase at ESTOC of 0.5– 0.6C from early morning
to the afternoon, which has been taken into account in the
harmonic formula. On annual timescales, net CO2 fluxes on
ESTOC were positive (outgassing) with an average value of
179 mmol CO2 m2. This value is 63% higher if we consider
a 6-hourly sampling frequency of wind instead of a monthly
frequency. This reinforces the previous finding by Bates and
Merlivat [2001] on the influence of short-term wind
variability on air-sea CO2 exchange. This reading was
generally found to be constant, except for 1999 where an
ingassing of 340 mmol CO2 m2 was obtained, relating to
the lowest temperature found during deep winter mixing
layer formation and a maximum of chlorophyll for ESTOC
during these 5 years. The highest net annual flux was found
in 1997 and 1998: 544 and 425 mmol CO2 m2, respectively, associated with the higher SST in summer time. With
an average annual wind speed of 7 m s1, a representative
figure for the area, this area would behave as a sink of CO2
of 0.15 mol m2 yr1. Takahashi et al.’s [2002]
calculations for this area for the year 1995 estimated a net
FCO2 ranging between 0 and 1 mol m2 yr1, considering
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Table 2. Annual Cycle of Average f CO2 at ESTOC Subdivided
Into Distinct Seasonal Time Period With Different Temperaturef CO2 Relationship
Period
Full year
March – October
October – January
January – March
2-9
upper mixed layer. The variation in CT between two series
of measurements, DCT, then
DCT ¼ DCT;exch þ DCT;mix þ DCT;bio :
ð4Þ
Relationship
f CO2
f CO2
f CO2
f CO2
=
=
=
=
101.5(±11.4) + 12.41 (0.55) T r2 = 0.893
112.2(±10.1) + 12.12 (0.48) T r2 = 0.962
42.18(±11.4) + 14.86 (0.52) T r2 = 0.981
369.3(±73) + 3.86 (3.8) T r2 = 0.371
the CO2 gas transfer coefficient of Wanninkhof [1992] and
climatological wind data. In order to improve our knowledge
and the effect of short-term wind variability, a continuous
and autonomous SAMI-CO2 system and a meteorological
station are going to be deployed in the ESTOC site, under the
European project ANIMATE.
3.4. Annual Cycle of Average Sea Surface f CO2
(Matm) Versus SST (°C) for the ESTOC Station
[23] Equation (3) revealed that surface CO2 largely
responded to temperature changes over the different
seasonal timescales. The linear regression slope for all
data was 12.4 matm C1 with a root-mean square
standard deviation (rms) around the regression line of
±7.22 matm. (Table 2). Bates et al. [1996, 1998] found that
the annual cycle of pCO2 in the Sargasso Sea could be
subdivided into different seasonal time periods with
different temperature-pCO2 relationships. The annual cycle
of the average fCO2 as a function of the average SST for the
ESTOC station is shown in Figure 5. From a thermodynamic perspective starting from an fCO2 equal to 339
matm (March 1997) the temperature increase, at a constant
CT (2101 mmol kg1), would lead to an fCO2 increase of
15.02 matm C1. However, between March and September,
the slope is only 12.12 matm C1 as a consequence of the
CT decrease, due first to biological activity and second to
the CO2 outflux. The thermodynamic conditions during the
cooling period, from October onward, with a lower CT
concentration (2096 mmol Kg1) and a temperature of
24.1C, would produce an fCO2 decrease of 14.5 matm
C1. From January to March, the f CO2 value increases due
to winter mixing. The corresponding linear regressions and
the rms errors for the three seasonal periods are presented in
Table 2. The rms errors are now much smaller (±3.10 matm
during autumn-early winter and ±4.28 matm during springsummer). During the mixing period in winter, the rms errors
were slightly lower than the annual relationship. These
temporal relationships of fCO2 and equation (3) can be used
to predict the surface fCO2 from satellite observations or
surface mooring.
3.5. Total CT Variability (DCT) and Contributions
From the Air-Sea Flux (DFCO2) and Biological and
Mixing Pumps in the Mixed Layer From 1996 –2000
[24] In order to establish the relative importance of the airsea flux and the biological pump in CT variability over the
period studied, total CT variation was calculated as the
difference between 2 months expressed in mmol m3,
(DCT = m2 m1). Seawater density is therefore also taken
into account. Positive values mean an increase of CT in the
The variation of carbon due to air-sea CO2 exchange,
DCT,exch, was assessed from the monthly averages of FCO2
determined each 6 hours and the average value of the
mixed layer (h1/2) between the two consecutive months
(DFCO2/h1/2). DCT,mix, the changes in CT due to mixing,
can be estimated from CT distribution and from the
physical structure of the mixed layer. The evaluation of
changes in CT due to biological processes can be obtained
from the difference DCT -DCT,exch-DCT,mix.
[25] Salinity variations in surface waters around ESTOC
site can largely be explained by a balance between vertical
exchange and surface evaporation/precipitation [Pelegrı́ et
al., 2003]. However, filaments originating at Cape Ghir in
the northwest African coastal upwelling area are a rather
ubiquitous feature and usually stretch offshore over 150 km
[Van Camp et al., 1991]. Following CZCS data and AVHRR
sensors, Davenport et al. [1999] suggested the Cape Ghir
filament should at least be considered a possible source of
advected trap material in traps deployed at 3000 m depth.
These authors also indicate that higher chlorophyll values
around October in ESTOC could be the result of temporal
advection of surface water from the northeast African
upwelling area and local enhancements of production.
Salinity of the surface water at the ESTOC site varied
seasonally between 36.6 and 36.9. Vertical mixing, which
occurs principally in winter from October onward, can
contribute to this salinity variation since surface water is
mixed at that time with water at around 150 m and salinities
of 36.2 – 36.3. For this season, a linear relationship between
salinity and CT has been found for ESTOC between the base
of the mixed layer and the first 400 m with a negative rate of
50.36 (±3.04, r2 = 0.923) mmol kg1 psu1. A similar slope
was found on horizontal transepts in winter around the
ESTOC site during the CANIGO project [Santana-Casiano
et al., 2001; Pérez et al., 2001]. At the end of summer,
Figure 5. Seasonally differentiated f CO2-temperature
relationships considering the full set of data. Regression
lines for each period are those shown in Table 1.
2 - 10
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 6. Monthly variations of CT (white open bars) calculated as the difference between two
consecutive months and the contributions due to (a) air-sea exchange (shaded open bar), (b) variations
due to vertical diffusion (shaded dashed bar) and entrainment (black dashed bar), (c) variation due to
biological processes (shaded open bar). The monthly variations are expressed in mmol m3.
salinity variations may be also due in part to horizontal
advection. In order to estimate the role of such advection on
the CT concentration in surface water, the relationship
between CT and salinity in surface water in a section
between ESTOC and a station located close to the upwelling
African area in the vicinity of Cape Ghir carried out during
at the end of October 1999 was examined [Pelegrı́ et al.,
2003]. The CT and salinity variations are, on average, in a
ratio of 65 (±2.38, r2 = 0.872) mmol kg1 psu1, which is
only slightly higher than the CT/S ratio that should be
observed in presence of only evaporation or precipitation
processes (about 60). This result shows a small effect of
advection on the surface values for inorganic carbon at
ESTOC during this cruise. To take into account the CT
variations due to salinity variations, the CT values (in mmol
m3) were normalized to an average salinity of 36.8 using
the ratios DCT/DS = 50 mmol kg1 psu1 from November to
March and DCT/DS = 65 mmol kg1 psu1 from April to
October [Bégovic and Copin-Montégut, 2002].
[26] Following Bégovic and Copin-Montegut [2002], the
changes in CT due to mixing were determined, after this
normalization, as the entrainment of water into the surface
layer and the vertical diffusion across the lower boundary of
the mixed layer,
DCT;mix ¼ 1=hðdh=dtÞ CTf CT þ 1=hKz ðdCT =dzÞ;
ð5Þ
where h is the mixed-layer depth at the time t, Kz is the
vertical diffusion coefficient (from 3 to 8 m2 d1) at the base
of the mixed layer (determined according to the formula of
Denman and Gargett [1983]). According to this model, only
the deepening of the mixed-layer induces mixing with the
underlying waters ( = 0 when dh/dt 0). In our study, CTf is
the CT concentration for the month m1 at the depth of the
mixed layer for the month m2.
[27] Figure 6 shows the monthly variability of each
contributor on equations (4) and (5) for the 5 years period
studied. The monthly change in CT induced by air-sea
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
2 - 11
Figure 7. (a) CT variations in mmol m3 at ESTOC after normalization to the salinity of 36.8 using the
procedure explained in the text. The lines join the experimental CT values (dots). (b) Variations of CT
(mmol m3) due to air-sea exchange with atmosphere (DCT,exch), mixing processes at the base of the
mixed layer (DCT,mix), and biological production (DCT,bio) for the periods of CT increase and decrease.
exchange DCT,exch was important in summer when outgassing of CO2 took place and the mixed- layer depth reached
around 30– 40 m. During the ingassing period from December to May, the DCT,exch was low (lower than 1.5 mmol m3)
since the mixed layer was thick. During the period when CT
was decreasing (from March to October) (Figure 7) the airsea exchange also contributed to decreasing CT in surface
waters (Figure 7b) with values from 8 mmol m3 (2000) to
18 mmol m3 (1997). During the periods of CT increase
(from October to March), air-sea exchange also increased by
2 ± 1 mmol m3, representing only 3% (1997) to 10%
(1999) of the total CT increase (Figure 7b).
[28] Mixing processes always increase C T since CT
increases with depth. The diffusive flux was important at
the end of autumn when the density gradients were low and
the mixed layer was deep. The contribution due to the
entrainment of water started just when the surface CT began
to increase as a result of the descent of the mixed layer; this
contribution does not take place from March – April onward
(Figure 6b). The total changes in CT due to mixing when CT
was increasing from October to March (Figure 7b) were as
high as 34 ± 4 mmol m3, representing 130% of the
increase in CT in 1996 and 1997 and 107% in 1999. During
the period from March to December when CT was decreasing, the mixing processes increased CT by 11.5 ± 1.5 mmol
m3, decreasing the net effect of both air-sea exchange and
biological activity.
[29] The biological consumption/respiration of carbon
deduced from this model gives us an estimate of the real
values due to the errors associated with vertical mixing and
air-sea exchanges determinations. Moreover, lateral advec-
tion can be underestimated due to the lack of data. The
calculated biological changes give us a photosynthetic
production from February to August (to September in
2000) with the highest values during April. From November
to January, a higher respiration activity could account for
the positive CT increase of 3 ± 2 mmol m3. The total
biological carbon uptake during the CT decreasing period
from March to October was 25.5 ± 5.7 mmol m3 (1999
was not considered in this calculation). These values represent between 80% (1997 and 1998) and 110% (1996 and
2000) of the total CT decrease. When the CT increases from
October to March, a lower biological carbon uptake of 7 ± 2
mmol m3 is deduced from equation (4) (Figure 7b). These
values were close to 0.32 ± 0.06 mmol C m3 d1 over the
5 years, with the highest value, 0.41 mmol C m3 d1,
observed during April 1999. According to Neuer et al.
[2002], annually integrated net primary production (PP) at
the ESTOC station from 1996 to 1998 was 0.23 ± 0.05
mmol m3 d1, with the highest values in March and
April when PP was 0.4 mmol m3d1. Moreover, at
ESTOC, C-export was only 0.2 mol C m2 yr1 (Table 1),
resulting in an annually averaged export ratio of only 0.02
for the Canary Island station. This lower export production,
when interpreted in a new production context, is indicative
of a lower supply of new nutrients into the productive
euphotic zone. Neuer et al. [2002] also compare yearly
integrated PP and C-export for the two oligotrophic sites,
ESTOC and BATS, showing a slightly lower mean PP
values at the ESTOC while C-export was only about 20% of
the 1 mol C m2 yr1 observed at BATS. The lower input of
new nitrogen both by smaller mesoscale eddy activity and
2 - 12
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 8. Time series of surface properties at the ESTOC site from 1996 to 2000 (open circles) and data
after removing the seasonality variability (detrended data, solid circles in a per months value) for
(a) temperature (C), (b) pHT at 25C (in mmol kg1), (c) total dissolved inorganic carbon normalized to
the average salinity for ESTOC of 36.8 using the procedure explained in the text, NCT (mmol kg1), and
(d) sea surface f CO2 (matm).
lack of significant nitrogen fixation at ESTOC compared
with BATS [McGillicuddy et al., 1999] are concluded from
their results. These differences, together with differing
temperature seasonal variability and wind regime, could
also explain the observed differences in the carbon system
variables between these two stations at both sites of the
same subtropical Atlantic gyre. In order to improve our
observation of the carbon cycles, further projects are
planned. These include the use of a SOMMA-type system
directly to determine CT in the water column and the
determination of primary productivity and 13C and 14C
together with several cruises from ESTOC to the upwelling
area in order clearly to describe the effect of filaments and
horizontal advection on the carbon cycle at the ESTOC site.
3.6. Trends in the Carbon Dioxide Parameters
[30] The variability of pH, CT, and f CO2 in the area is the
result of both the annual heat pattern and the strong wind
forces which regulate biological production in the area.
While surface alkalinity has a seasonal variability of
25 mmol kg1 over the 5-year observation period, the
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
NAT ESTOC data showed scant variability from the mean
of 2291.5 ± 2.5 mmol kg1. From 1996 to 2000, surface AT
did not show any noticeable increase which might contribute to the observed f CO2. In order to remove seasonal
variability in the sea surface temperature, the pHT, NCT, and
f CO2 data were treated using the SPSS program (version
10.0) and the routine trend cycle (Figure 8). From 1996 to
2000, detrended NCT (at salinity S = 36.8) concentration
measured at the Canary Island station showed an annual
increase of 0.39 ± 1.6 mmol kg1 yr1. A larger error of
±9.1 mmol kg1 yr1 in nondetrended NCT reflects the
seasonal variability of surface CT. The BATS station shows
that from 1988 to 1998, surface NCT increased at a rate of
1.6 ± 5.8 mmol kg1 yr1 after seasonal variability was
removed (Table 1). However, when the data corresponding
to 1993 – 1998 (a 5-year period, as in our study) were
considered, a much smaller increase in NCT could be
inferred [Bates, 2001; personal communication, 2003].
Detrended sea surface temperature does not follow any
noticeable trend. What can be observed is that the highest
temperatures over these 5 years were recorded in 1997,
when the detrended values increased 0.95C with respect to
1996. Moreover, our 5 years of observations show an
increase for f CO2 in surface seawater. If seasonal detrended
f CO2 data are considered, seawater f CO2 increases at a rate
of 0.71 ± 5.1 matm yr1. This increase should theoretically
produce in itself a 0.42 mmol kg1 yr1 increase in NCT,
similar to the observed detrended value. As was expected,
pHt-25C is affected by the increase in f CO2, decreasing
over these 5 years at a rate of 0.0015 pH units yr1.
Relatively small annual trends and higher associated errors
can be explained if we bear in mind that the sampling period
is not long enough adequately to resolve a rate of NAT and
NCT increase, given the errors associated with both
seasonally detrended or nondetrended data. Similarly, no
interannual variability of hydrographic properties during
this period was discerned. From a statistical point of view, a
series of 10 years of observations is necessary adequately to
describe the interannual trends of CO2 concentration in the
Northeast Atlantic subtropical gyre, as is the case at the
BATS and ALOHA sites [Bates, 2001; Winn et al., 1998].
4. Conclusions
[31] From the data obtained during the time period
between 1996 and 2000, we can conclude that at the
ESTOC station, sited in the subtropical east Atlantic, a
quasi-equilibrium is achieved by a balance between late
winter CO2 uptake and late summer outgassing. However,
the predominant trade winds blowing in this area during the
summer time favor CO2 outgassing at a rate of 0.18 mol m2
yr1. During this period, from March to October, CT
decreases (30 – 35 mmol m3). Mixing at the base of the
mixed layer (11.5 ± 1.5 mmol m3) is compensated by the
air-sea exchange, while a net organic production of about
25.5 ± 5.7 mmol m3 is estimated, constituting between
80% (1997 and 1998) and 110 % (1996 and 2000) of the
total CT decrease. On an annual scale, mixing processes
control the CT increase from October to the end of autumn,
while biological drawdown accounts for the decrease in
2 - 13
inorganic carbon from March to October. A relatively low
f CO2 increase, 0.71 ± 5.1 matm yr1, after removing
seasonal variability, was determined after 5 years of study at
the ESTOC, as compared to the 1.4 ± 10.7 matm yr1
increase observed in the 10-year study at BATS. Increased
temperature and salinity seasonal variability and wind
regimes, together with higher input of new nitrogen both
by smaller mesoscale eddy and nitrogen fixation at BATS
with respect to ESTOC, may account for the observed
differences between both stations of the subtropical North
Atlantic. Ten years of study are now planned to improve our
knowledge of the role of the processes controlling carbon
dioxide variability in the open ocean time series station
ESTOC.
[32] Acknowledgments. Part of this work has been supported by
the European Commission, through the MAST III Programme, and the
CANIGO project (MAS3-CT96-0060). Magdalena and Melchor wish to
thank F. J. Millero, who introduced them to the study of the CO2 system to
be applied in the Canary Island area. We also thank N. Bates for his
comments and recommendations. The authors thank all the participants in
the cruises, together with the masters, officers and crew of the research
vessels Taliarte, Meteor, Victor Hensen, and Poseidon. Our grateful thanks
go to the dedicated ULPGC technicians, Victor Siruela and Pedro Carmona,
who automated the pH system and carried out pCO2 control programs. We
would also like to thank Miguel Laglera, Ma Jesús, and Ivan for helping to
collect CO2 samples on some cruises and Heather Adams for correcting the
English version. This paper benefited from two anonymous reviewers.
References
Abraham, B., and J. Ledolter, Statistical Methods for Forecasting, John
Wiley, New York, 1983.
Bates, N. R., Interannual variability of oceanic CO2 and biogeochemical
properties in the western North Atlantic subtropical gyre, Deep Sea Res.,
Part II, 48, 1507 – 1528, 2001.
Bates, N. R., and L. Merlivat, The influence of short-term wind variability
on air-sea CO2 exchange, Geophys. Res. Lett., 28, 3281 – 3284, 2001.
Bates, N. R., A. F. Michaels, and A. H. Knap, Seasonal and interannual
variability of oceanic carbon dioxide species at the U. S. JGOFS Bermuda
Atlantic Time-series Study (BATS) site, Deep Sea Res., Part II, 43, 347 –
383, 1996.
Bates, N. R., T. Takahashi, D. W. Chipman, and A. H. Knap, Variability
of pCO2 on diel to seasonal timescales in the Sargasso Sea near
Bermuda, J. Geophys. Res., 103, 15,567 – 15,585, 1998.
Bégovic, M., and C. Copin-Montégut, Processes controlling annual variations in the partial pressure of CO2 in surface waters of the central northwestern Mediterranean Sea (Dyfamed site), Deep Sea Res., Part II, 49,
2031 – 2047, 2002.
Bellerby, R. G. J., D. R. Turner, G. E. Millward, and P. J. Worsfold, Shipboard flow injection determination of sea water pH with spectrophotometric detection, Anal. Chim. Acta, 309, 259 – 270, 1995.
Brainerd, K. E., and M. C. Greg, Surface mixed and mixing layer depths,
Deep Sea Res., Part I, 42, 1521 – 1543, 1995.
Broecker, W. S., and T. H. Peng, Tracers in the Sea, 690 pp., Eldigio, New
York, 1982.
Butler, J. H., C. M. Elkins, C. M. Bruson, K. B. Egan, T. M. Thompson, T. J.
Conway, and B. D. Hall, Trace gases in and over the west Pacific and east
Indian Oceans during the El Niño-Southern Oscillation event of 1987,
NOAA Data Rep. ERL-ARL-16, 104 pp., Air Resour. Lab., Silver Spring,
Md., 1988.
Clayton, T. D., and R. H. Byrne, Spectrophotometric seawater pH measurements: Total hydrogen ion concentration scale calibration of m-cresol
purple and at-sea results, Deep Sea Res., Part I, 42, 411 – 429, 1993.
Davenport, R., S. Neuer, A. Hernández-Guerra, M.-J. Rueda, O. Llinás,
G. Fisher, and G. Wefer, Seasonal and interannual pigment concentration
in Canary Islands region from CZCS data and comparison with observations from ESTOC, Int. J. Remote Sens., 20, 1419 – 1433, 1999.
Davenport, R., S. Neuer, P. Helmke, J. Pérez-Marrero, and O. Llinás, Primary productivity in the northern Canary Islands region as inferred from
SeaWiFS imagery, Deep Sea Res., Part II, 49, 3481 – 3496, 2002.
Denman, K. L., and A. Gargett, Time and space scales of mixing and
advection of phytoplankton in the upper ocean, Limnol. Oceanogr., 28,
801 – 815, 1983.
2 - 14
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Dickson, A. G., An exact definition of total alkalinity and a procedure for
the estimation of alkalinity and total inorganic carbon from titration data,
Deep Sea Res., 28, 609 – 623, 1981.
Dickson, A. G., , and C. Goyet (Eds.), Handbook of methods for the
analysis of the various parameters of the carbon dioxide system in sea
water, Rep. ORNL/CDIAC-74, 180 pp., U.S. Dep. of Energy, Washington,
D. C., 1994.
Dickson, A. G., and F. J. Millero, A comparison of the equilibrium constants for the dissociation of carbonic acid in seawater media, Deep Sea
Res., 34, 1733 – 1743, 1987.
González-Dávila, M., J. M. Santana-Casiano, and L. M. Laglera, Report
and preliminary results of Poseidon cruise 237/2, Vigo-Las Palmas, Ber.
Fachber. Geowiss.121, pp. 19 – 22, Univ. of Bremen, Bremen, Germany,
1998.
González-Dávila, M., J. M. Santana-Casiano, D. de Armas, J. Escanez, and
P. Guillén, The influence of island generated eddies on the carbon dioxide
system distribution south of the Canary Islands, J. Mar. Syst, in press,
2003.
Hansen, H. P., and K. Grasshoff, Automated chemical analysis, in Methods
of Seawater Analysis, edited by K. Grasshoff, M. Ehrhardt, and
K. Kremling, pp. 347 – 395, Verlag Chemie, Weinheim, Germany, 1983.
Karl, D. M., J. E. Dore, R. Lukas, A. F. Michaels, N. R. Bates, and A. Knap,
Building the long-term picture: The U.S. JGOFS time-series programs,
Oceanography, 14, 6 – 17, 2001.
Lee, K., R. Wanninkhof, R. A. Feely, F. J. Millero, and T. H. Peng, Global
relationships of total inorganic carbon with temperature and nitrate in
surface water, Global Biogeochem. Cycles, 14, 973 – 994, 2000.
Llinás, O., A. Rodrı́guez de León, G. Siedler, and G. Wefer, ESTOC data
report 1994, Inf. Téc. 3, 72 pp., Inst. Canario Cienc. de Mar., Telde,
Canary Islands, 1997.
Llinás, O., A. Rodrı́guez de León, G. Siedler, and G. Wefer, ESTOC data
report 1995 – 1996, Inf. Téc. 7, 152 pp., Inst. Canario Cienc. de Mar.,
Telde, Canary Islands, 1999.
McGillicuddy, D. J., R. Johnson, D. A. Siegel, A. F. Michaels, N. R. Bates,
and A. H. Knap, Mesoscale variations of biogeochemical properties in
the Sargasso Sea, J. Geophys. Res., 104, 13,381 – 13,394, 1999.
Michaels, A. F., and A. H. Knap, Overview of the U.S. JGOFS Bermuda
Atlantic Time-series Study and the Hydrostation S program, Deep Sea
Res., Part II, 43, 157 – 198, 1996.
Michaels, A. F., et al., Seasonal patterns of ocean biogeochemistry at the
U.S. JGOFS Bermuda Atlantic Time-series Study site, Deep Sea Res.,
Part I, 41, 1013 – 1038, 1994.
Millero, F. J., K. Lee, and M. Roche, Distribution of alkalinity in the surface
waters of the major oceans, Mar. Chem., 60, 95 – 110, 1998.
Mintrop, L., F. F. Pérez, M. González-Dávila, J. M. Santana-Casiano, and
A. Körtzinger, Alkalinity determination by potentiometry: Intercalibration using three different methods, Cienc. Mar., 26, 23 – 37, 2000.
Neuer, S., R. Davenport, T. Freudenthal, G. Wefer, O. Llinás, M.-J. Rueda,
D. K. Steinberg, and D. M. Karl, Differences in the biological carbon
pump at three subtropical ocean sites, Geophys. Res. Lett., 29(18), 1885,
doi:10.1029/2002GL015393, 2002.
Pelegrı́, J. L., J. Arı́stegui, L. Cana, M. González-Dávila, A. HernándezGuerra, S. Hernández-León, A. Marrero-Dı́az, M. F. Montero, P. Sangrá,
and M. Santana-Casiano, Coupling between the open ocean and the
coastal upwelling region off northwest Africa: Water recirculation and
offshore pumping of organic matter, J. Mar. Syst, in press, 2003.
Pérez, F. F., L. Mintrop, O. Llinás, M. González-Dávila, C. G. Castro,
M. Alvarez, A. Körtzinger, J. Santana-Casiano, M.-J. Rueda, and A. F.
Rios, Mixing analysis of nutrients, oxygen and inorganic carbon in the
Canary Islands region, J. Mar. Syst., 28, 183 – 201, 2001.
Pérez-Marrero, J., Aplicación de la radiometrı́a infrarroja a las observaciones oceanográficas en el Atlántico Centro Oriental, Ph.D thesis, 198
pp., Phys. Dep., Univ. of Las Palmas, Las Palmas de Gran Canarias,
Spain, 1998.
Santana-Casiano, J. M., M. González-Dávila, L. M. Laglera-Baquer, and
M. J. Rodrı́guez-Somoza, Carbon dioxide system in the Canary region
during October 1995, Sci. Mar., 65, 41 – 49, 2001.
Takahashi, T., J. Olafsson, J. G. Goddard, D. W. Chipman, and S. C.
Sutherland, Seasonal variation of CO2 and nutrients in the high-latitude
surface oceans: A comparative study, Global Biogeochem. Cycles, 7,
843 – 878, 1993.
Takahashi, T., et al., Global sea-air CO2 flux based on climatological surface ocean pCO2 and seasonal biological and temperature effects, Deep
Sea Res., Part II, 49, 1601 – 1622, 2002.
Tans, P. P., I. Y. Fung, and T. Takahashi, Observational constraints on the
global atmospheric CO2-budget, Science, 247, 1431 – 1438, 1990.
Van Camp, L., L. Nykjaer, E. Mittelstaedt, and P. Schlittenhardt, Upwelling and boundary circulation off Northwest Africa as depicted by
infrared and visible satellite observations, Prog. Oceanogr., 26, 357 –
402, 1991.
Wanninkhof, R., Relationship between wind speed and gas exchange over
the ocean, J. Geophys. Res., 97, 7373 – 7382, 1992.
Wanninkhof, R., and R. A. Feely, fCO2 dynamics in the Atlantic, South
Pacific and south Indian oceans, Mar. Chem., 60, 15 – 31, 1998.
Wanninkhof, R., and W. R. McGillis, A cubic relationship between air-sea
CO2 exchange and wind speed, Geophys. Res. Lett., 26, 1889 – 1892,
1999.
Wanninkhof, R., and K. Thoning, Measurement of fugacity of CO2 in surface water using continuous and discrete sampling methods, Mar. Chem.,
44, 189 – 204, 1993.
Welschmeyer, N. A., Fluorometric analysis of chlorophyll a in presence of
chlorophyll b and phaeopigments, Limnol. Oceanogr., 39, 1985 – 1992,
1994.
Winn, C. D., F. T. Mackenzie, C. J. Carrillo, C. L. Sabine, and D. M. Karl,
Air-sea carbon dioxide exchange in the North Pacific Subtropical Gyre:
Implications for the global carbon budget, Global Biogeochem. Cycles, 8,
157 – 163, 1994.
Winn, C. D., Y.-H. Li, F. T. Mackenzie, and D. M. Karl, Rising surface
ocean dissolved inorganic carbon at the Hawaii ocean-time series site,
Mar. Chem., 60, 33 – 47, 1998.
World Ocean Circulation Experiment (WOCE), WOCE operations manual,
vol. 3, sect. 3.1, part 3.1.3: WHP operations and methods, WHP Off. Rep.
WHPO 91-1, WOCE Rep. 68/91, 12 pp., Woods Hole, Mass., 1994.
E.-F. González-Dávila, Statistical Department, Faculty of Mathematics,
University of La Laguna, Spain. ([email protected])
M. González-Dávila and J. M. Santana-Casiano, Chemistry Department,
Faculty of Marine Science, University of Las Palmas de Gran Canaria,
35017 Las Palmas, Gran Canarias, Spain. ([email protected];
[email protected])
O. Llinás and M.-J. Rueda, Instituto Canario de Ciencias Marinas,
Gobierno de Canarias, Taliarte, Telde, Gran Canaria, Spain. (ollinas@
iccm.rcanaria.es)
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 1. Sea surface temperature from the Advanced Very High Resolution Radiometer (AVHRR)
sensor of the National Oceanic and Atmospheric Administration (NOAA) for the Canary Islands area,
showing the position of the ESTOC site for November 10, 1999.
2-4
GONZÁLEZ-DÁVILA ET AL.: SEA-SURFACE CARBON DIOXIDE AT ESTOC
Figure 2. Distribution of (a) temperature and chlorophyll a concentration (mg m3) for the first 200 m
on ESTOC from 1996 to 2000 and (b) salinity and nitrate + nitrite (N + N) (mmol kg1) for two
representative years, 1996 and 1997.
2-5