Marine Chemistry 93 (2005) 81 – 103

www.elsevier.com/locate/marchem

Iron, macronutrients and diatom blooms in the Peru upwelling

regime: brown and blue waters of Peru
Kenneth W. Brulanda,*, Eden L. Ruea, Geoffrey J. Smitha, Giacomo R. DiTulliob
a
Institute of Marine Sciences and Department of Ocean Sciences, University of California at Santa Cruz, CA 95064, United States
b
Hollings Marine Lab, College of Charleston, Charleston, SC 29412, United States

Received 24 November 2003; received in revised form 4 June 2004; accepted 23 June 2004
Available online 18 December 2004

Abstract

Surface water transects and vertical profiles for dissolved iron, macronutrients, chlorophyll a (Chl a), and hydrographic data
were obtained in the Peru upwelling regime during August and September 2000. The supply of the micronutrient iron, relative
to that of the macronutrients nitrate, phosphate and silicic acid, is shown to play a critical role in allowing extensive diatom
blooms to develop in the Peru upwelling system. The extremely high-chlorophyll bbrown waters of PeruQ (with Chl a
concentrations between 20 and 45 Ag/l) result from massive diatom blooms with maximal photochemical efficiencies
( F v/F m N 0.6) occurring in the iron-rich upwelling region observed over the broad continental shelf off northern and central
Peru. The source of the upwelled water in this region is the nutrient-rich subsurface countercurrent in contact with the organic-
rich shelf sediments. This subsurface shelf water is suboxic and has extremely high concentrations of dissolved Fe (N50 nM) in
the near-bottom waters. In marked contrast, relatively low-chlorophyll bblue watersQ (Chl a b 2 Ag/l) with low concentrations of
dissolved Fe (b0.1 nM) and high unutilized macronutrient concentrations are observed in the coastal upwelled waters along the
southern coast of Peru and in the offshore regions of the Peru Current. Southern Peru is a region without a wide shelf to serve as
a source of iron and, as a result, dissolved Fe concentrations in the near-bottom suboxic waters of this region are an order-of-
magnitude lower than observed off northern and central Peru. In addition, the offshore Peru Current is a broad, Fe-limited, high-
nitrate, lower than expected chlorophyll region extending hundreds of kilometers offshore into the northeast region of the South
Pacific subtropical gyre and northwestward into the South Equatorial Pacific.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Chlorophyll a; Near-bottom suboxic waters; Maximal photochemical efficiencies

1. Introduction

The biological richness of the Peruvian coastal

* Corresponding author. Tel.: +1 831 459 4587. waters is legendary (Brink et al., 1983). Southeasterly
E-mail address: bruland@ucsc.edu (K.W. Bruland). trade winds along the coast of Peru result in wind-
0304-4203/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.marchem.2004.06.011
82 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
driven coastal upwelling of colder, nutrient-rich
subsurface water. Intense upwelling of the macro-
nutrients nitrate, phosphate, and silicic acid to surface
waters in this region allows extensive phytoplankton
blooms to occur that may extend tens to hundreds of
kilometers offshore. Large diatoms tend to dominate
the biomass in phytoplankton blooms that develop in
these coastal upwelling regimes (de Mendiola, 1981;
Wilkerson et al., 2000), and it has been argued that
diatom-driven new production efficiently fuels the
food chains leading to fish production (Smetacek,
1998). With its year-round, persistent, coastal upwell-
ing, the Peru Current system is one of the most
productive marine ecosystems (Barber and Smith,
1981; Carr, 2002), supporting intensive pelagic fish-
eries, most notably the anchoveta (Engraulis ringens;
Alheit and Bernal, 1993).
With this rich productivity, it is surprising that
high-nitrate, lower than expected chlorophyll (HNLC)
conditions have been reported for the Peru upwelling
regime. Thomas (1972, 1979) in some of the first
discussions of HNLC regimes pointed out that blow-
Chl, high-nutrient water is not unique to offshore
regionsQ and that bOff Peru, dblueT water containing 2
Ag Chl/l and 10 Ag—at NO3 —N/L was found
(Strickland et al., 1969).Q These authors found it
bremarkableQ that offshore waters of the Peru Current
had not been depleted of nutrients by phytoplankton
assimilation. The grazing hypothesis (Walsh, 1976)
has been generally accepted as the principle explan-
ation for this observation. It has been suggested that
zooplankton are more abundant in the euphotic zone
off Peru because of the suboxic water just below, and
that this results in more intensive grazing pressure at
the beginning of the bloom, keeping the phytoplank-
ton growth under control from the outset (Minas and
Minas, 1992). Hutchins et al. (2002), however,
recently investigated phytoplankton iron limitation
in the high-nutrient Peru Current system and demon-
strated that primary producers can be limited by lack
of iron in this regime.
It has recently been demonstrated that extensive
parts of the California upwelling system are iron
limited (Hutchins and Bruland, 1998; Hutchins et al.,
1998; Bruland et al., 2001; Firme et al., 2003).
Coastal diatoms have high iron requirements (Sunda
and Huntsman, 1995; Bruland et al., 2001) that cannot
be met by the small amount of iron present with the
macronutrients in subsurface oceanic waters (Bruland
et al., 2001). The presence or absence of a broad
continental shelf can greatly influence the external
supply of iron to the upwelled surface waters in
coastal upwelling regions (Johnson et al., 1999, 2001;
Bruland et al., 2001). Coastal upwelling off central
and northern California is most intense in the spring
and summer months and this peak in upwelling is
temporally out of phase with the river input occurring
during the winter months in this semiarid climate.
During periods of intense upwelling over a broad
shelf, however, the interaction of upwelling water
with the shelf sediments can transport elevated
concentrations of iron to the surface waters. When
upwelling is focused offshore of a narrow continental
shelf, the upwelled waters are characterized by low
dissolved and particulate iron concentrations, and
these waters rapidly develop into coastal upwelling
HNLC systems limited by the micronutrient iron
(Bruland et al., 2001). Supplemental iron additions to
these Fe-limited systems in shipboard grow-out
experiments resulted in nitrate drawdown and pro-
duced blooms of large diatoms (Hutchins and Bru-
land, 1998; Hutchins et al., 1998; Firme et al., 2003).
The same physical characteristics that appear to be
predictors of Fe-limited areas along the central and
northern California coast—intense coastal upwelling,
together with a narrow continental shelf and negli-
gible fluvial and eolian inputs of iron—are present in
some areas of the Peru upwelling regime. The Peru
coast is part of an active continental margin with a
relatively narrow shelf, a steep slope and an oceanic
trench located just offshore. The width of the
continental shelf along Peru varies from 150 km off
northern and central Peru to less than 10 km off parts
of southern Peru (Fig. 1B). As a result, we predict a
mosaic of conditions ranging from iron-replete to
iron-limited such as observed off central California
(Hutchins et al., 1998; Bruland et al., 2001; Firme et
al., 2003). This expectation is supported by historical
observations of extremely high Chl a concentrations
over the northern Peru inner shelf, referred to as the
bbrown waters of PeruQ by Strickland et al. (1969) and
by observations of extensive bblue waterQ regions
exhibiting HNLC conditions in the Peru upwelling
region (Minas and Minas, 1992).
Coastal Peru has a hyperarid climate with the
landscape dominated by remarkably barren hills and
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Fig. 1. The locations of surface transects (A) and the bathymetry (B) off the coast of Peru.
the Atacama Desert. What little rainfall there is occurs
almost exclusively in the Andes and not in the coastal
range (Hebbeln et al., 2000; Scholl et al., 1970). The
maximum discharge of this precipitation via coastal
rivers occurs in the austral summer months of January
to April as a result of snowmelt and rainfall at high
elevations along the western flank of the Andes
(Scheidegger and Krissek, 1981). Although substan-
tial upwelling in the Peru area occurs year-round, the
maximum coastal upwelling occurs from June through
November (Scheidegger and Krissek, 1981) because
of the maximum in southeast trade wind intensities
(Bakun, 1987; Bakun and Nelson, 1991). During this
period, however, negligible fresh water or terrigenous
sediment is introduced directly into the coastal marine
environment off Peru. We planned our investigation to
take place in August/September—during the max-
imum upwelling period and during a time with
negligible direct river input.
Eolian input of iron to the upwelling region off the
coast of Peru is low. The southeasterly trade winds
remain relatively constant in direction and thus deliver
little of the dust from the arid Peru continent to the
offshore waters. This constancy in the direction of the
winds off the coast of Peru is a peculiarity noted as
early as 1553 by Cieza de Leon when he noted bThere
83
is another curious thing, which is that there is only one
wind on this coast, and that is from the southQ (Brink
et al., 1983). Fung et al. (2000) have modeled both the
amount of new dissolved iron supplied via eolian
deposition and the iron assimilation required for the
amount of new production. The iron delivered from
the eolian flux of dust, off the coast of Peru accounts
for less than a few percent of the iron required for the
observed productivity in this region. Moore et al.
(2002) have similarly modeled nutrient-limitation
patterns and the majority of the Peru Current system
is an area with low eolian iron input and is suggested
to be iron-limited.
The poleward flowing, subsurface Peru under-
current is the major source of water upwelling along
the Peru coast (Nelson and Neshyba, 1979; Brink et
al., 1983). This nutrient rich undercurrent is contin-
uous from 58S to at least 158S, flowing just below the
surface layer over the continental shelf and inner
slope. The undercurrent in this region is a suboxic
water mass (with evidence of substantial denitrifica-
tion—but generally not sulfate reduction) in contact
with a mud lens of organic rich sediments (which
become anaerobic near the sediment water interface)
located on the shelf. This organic-rich mud lens,
starting at depths of roughly 50 m and extending
84 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
offshore to just past the shelf break (Scheidegger and
Krissek, 1981), is the major site for the deposition of
fine-grained terrigenous sediments transported to this
area by coastal streams and rivers during the southern
hemisphere summer months. It is also the major
deposition site for significant amounts of the export
production of biogenic organic matter from the
overlying water column.
We predicted that the suboxic condition of the
waters in contact with these organic rich shelf sedi-
ments would result in high concentrations of dissolved
iron that could be supplied along with macronutrients
during coastal upwelling occurring over these rela-
tively broad shelf areas off northern and central Peru.
Literature evidence in support of this idea is provided
by iron data from Hong and Kester (1986). They
observed high concentrations of dissolved Fe (N50
nM) in the suboxic near bottom water over the Peru
shelf at depths of 40 to 100 m within 5 to 10 km of
shore at latitudes of 108S to 118S. Thus, it appears the
suboxic undercurrent in contact with shelf sediments,
particularly in the northern and central regions of Peru
with its relatively broad shelf, may be a particularly
rich source of iron to complement the macronutrients
in upwelled waters in these regions.
Herein we present surface transect and vertical
profiles from the Peru upwelling region that provide
spatial coverage of dissolved Fe, macronutrients, Chl
a, and hydrographic data. These data sets provide
evidence for the importance of iron in influencing
phytoplankton bloom dynamics in the Peru upwelling
system. In a complementary paper, DiTullio et al.
(submitted for publication) present transects and
vertical profiles of biological parameters such as
flavodoxin/ferredoxin, photosynthetic competency
( F v/F m), and algal taxonomic composition based on
pigment data that complement our chemical data in
supporting the important role of iron. A small subset
of the data from DiTullio et al. (submitted for
publication) will be included here to facilitate direct
comparison with the dissolved Fe measurements. In a
recent paper, Hutchins et al. (2002) presented ship-
board incubation experiments from two stations
within the Peru HNLC bblue waterQ region. Results
from varying iron additions confirmed that primary
producers in part of the Peru upwelling system were
indeed limited by iron. This observation was con-
firmed by a study by Eldridge et al. (2004).
2. Methods
Oceanographic data was collected in the eastern
tropical Pacific between 18 August and 28 September
2000 aboard the R/V Melville. Hydrographic data for
vertical profiles were collected using a Seabird CTD
system equipped with a Seabird dissolved oxygen
sensor, a fluorometer and a transmissometer. Under-
way sea surface temperature and salinity measure-
ments were monitored from the ship’s flow-through
seawater system in conjunction with other meteoro-
logical parameters using the R/V Melville’s shipboard
bIMETQ data acquisition system.
Samples for the determination of iron were
collected by two different methods. Subsurface
samples were collected by use of 30-l Teflonk-
coated GO-Flok bottles (General Oceanics) attached
to Kevlark hydroline (Bruland et al., 1979). Surface
samples were collected with a clean surface pump
bsipperQ system using an all PTFE Teflonk dia-
phragm pump (Bruiserk, Osmonics) and PFA Tef-
lonk tubing mounted to a PVC depressor vane 1 m
above a 20-kg PVC fish attached to Yacht-Braidk
dacron line (New England Rope). For underway
surface sampling at speeds from 1 to 7 knots, the
sipper system was deployed off the side of the ship
using the ship’s crane to suspend the fish outside of
the wake with the intake at approximately 2-m deep.
Faster speeds are possible with this sipper design if
there is little or no swell and the sipper remains
outside of any breaking bow waves. The sipper design
also allowed near-stationary sampling (moving for-
ward into clean water at 0.5 to 1 knots) in order to
collect large volumes of trace–metal–clean seawater at
depths up to 15 m.
Samples were filtered through either acid-cleaned
142-mm diameter 0.4-Am absolute pore size poly-
carbonate track-etched (PCTE) membrane filters
(Nucleporek) held in PTFE Teflonk filter sand-
wiches (Millipore) or through acid-cleaned 0.45-Am
absolute pore size Teflonk membrane polypropylene
capsule filters (Calyxk, MSI). The absolute pore
size Teflonk membrane capsule filters were chosen
to address previous concerns about possible filtration
artifacts from the use of woven polypropylene depth–
filter capsules with a nominal 0.45-Am cutoff. No
significant differences in the bdissolvedQ metal
concentrations were observed between the PCTE
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
membrane and capsule filtering methods (data not
shown).
The conditions used for cleaning were different for
both types of filters and were as follows: PCTE filters
were individually rinsed with Milli-Qk H2O and
placed in a sealed plastic container to allow sequential
soakings with either HCl or HNO3 over a period of
several weeks [6 M HCl for 2 weeks, 2 M HNO3 for 2
weeks, 3 M HCl for 3 weeks, 4 M HNO3 for 1 week,
1 M Q-HCL (quartz subboiled hydrochloric acid) for
1 week]. Care was taken to separate the filters in each
acid bath and during batch rinses to ensure complete
flushing of the solutions between each filter. The
filters were batch rinsed between acids with Milli-Qk
H2O until the pH of the rinse water remained pH 4.5–
5. After the last soaking with 1M Q-HCl, the filters
were individually rinsed with quartz subboiled,
distilled water (Q-H2O) and placed to dry in a class
100 clean bench in the same Q-H2O-rinsed plastic
container used to clean the filters. The dried filters
were placed individually into similarly acid-cleaned
polyethylene zipper bags until use. This rigorous
cleaning procedure allows for the subsequent diges-
tion of the filter and particulates for determination of
low-level suspended particulate trace metal concen-
trations from open ocean samples.
The Teflonk membrane polypropylene capsule
Calyxk (Osmonics) filters were cleaned by pumping
the different solutions through a string of eight capsule
filters connected inline with Bev-a-linek tubing and
polypropylene JACOk fittings. The capsules were
filled with Optimak grade (Fisher Scientific) meth-
anol to wet the hydrophobic Teflonk membrane then
emptied and rinsed with Milli-Qk H2O until all
methanol was removed. The filters were then filled
with 4 M HNO3 taking care to remove all air and
soaked for one week. After rinsing with Milli-Qk H2O
until the pH was 4.5–5, the filters were filled with 4 M
HCl taking care to remove all air and soaked for another
2 weeks. The filters were rinsed again with Milli-Qk
H2O until the pH was 4.5–5 and then individually
stored with Milli-Qk H2O inside until use.
Dissolved iron was measured by two different
methods. Dissolved Fe was determined at sea using
adsorptive cathodic stripping voltammetry (ACSV) on
samples that had been acidified to pH~1.7 with 6 M
Q-HCl and microwaved at 1100 W for 30 s
(approximately 50-ml samples microwaved twice for
85
15 s each and allowed to cool for 10 min; Bruland et
al., 2001; Bruland and Rue, 2001). This pretreatment
coupled with ACSV has been shown to provide an
accurate measurement of total dissolved Fe in samples
that originally were influenced by strong Fe-binding
organic ligands or colloids that bind the Fe under
ambient conditions. The acidified and microwaved
samples were then neutralized slowly using 10 Al
additions of 1 M Q-NH4OH, followed by the
immediate addition of 27.5 AM of SA as the
competing ligand and the buffer (Rue and Bruland,
1995, 1997). On a subset of samples, labile dissolved
iron was determined with no acidification or micro-
wave pretreatment prior to adding the buffer and 27.5
AM of SA. For these labile iron determinations, a 15-
min equilibration step was used prior to the ACSV
determination.
Dissolved Fe was also determined in our shore-
based clean-lab for seawater samples acidified for N6
months to a pH of 1.7–1.8 using 4 ml of 6 M Q-HCl/l
sea water. A modified PDC/DDC (pyrrolidine- and
diethyl-dithiocarbamate) organic solvent extraction
technique described in Bruland et al. (1979) was
used. The modified procedure uses 1 ml of 1.6% w/w
of PDC/DDC in 1% w/w Q-NH4OH (purified by
extracting 15 ml then 43 ml with Q-CHCl3) added
to 250 g of seawater then buffered to pH 4.3–4.5 with
saturated Q-ammonium acetate (Q-NH4Ac, cleaned in
a similar manner to the PDC/DDC) and extracted into
8 and 6 ml of Q-CHCl3. No nitric acid back extraction
was performed in order to minimize recovery losses
for cobalt and possible contamination during back-
extraction. Instead, both chloroform extracts were
combined in a 125-ml Teflonk separatory funnel
containing 5 ml of Q-H2O to rinse the organic phase
of any seawater salts carried over with the extracted
trace metals in the chloroform phase. The chloroform
was then drained into quartz beakers containing 100
Al of concentrated Q-HNO3 and gently evaporated
overnight on a hotplate in a class 100 clean-hood. The
residue was further oxidized to a slight amber color
with the addition and drying of two more 100 Al
aliquots of Q-HNO3 before transfer with three 500 Al
aliquots of 1 M Q-HNO3 into a preweighed 1.5 ml
polypropylene microcentrifuge tube. Extraction
blanks were prepared by reextracting previously
extracted seawater using exactly the same conditions
as described for the samples. Iron concentrations in
86 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
the organic extracts were measured using a Perkin-
Elmer 4100ZL graphite furnace atomic absorption
spectrometer (GFAAS) with THGA conditions and
Zeeman background correction. This method has been
well documented to provide an accurate measurement
of total dissolved Fe (Martin and Gordon, 1988;
Bruland and Rue, 2001).
A series of 12 samples collected by GO-Flok
bottles were analyzed by both our organic extraction/
GFAAS method back at UCSC and by our ACSV
method on board ship. We found excellent agreement
between the two methods and both appear to provide
an accurate measure of total dissolved Fe. When the
ACSV data was plotted against the organic extraction/
GFAAS data over the range of concentrations from
0.1 to N100 nM, the slope of the linear correlation was
0.973, with an r 2=0.996 (the average ratio of organic
extraction/GFAAS–ACSV was 1.03). A paired-data t-
test on the data set showed that the results of the two
analytical approaches were nondistinguishable at the
99% confidence interval. Both methods involve acid-
ifying the sample to pH 1.7 with HCl, and either
microwaving for shipboard analysis or storage for a
period of months for analysis back at UCSC. Both
methods determine total dissolved Fe, regardless of
the original speciation; they measure the sum of what
was originally present as Fe(II) and Fe(III); Fe
existing as inorganic or organic complexes, or as
colloidal Fe.
Nitrate (plus nitrite), phosphate, and silicic acid
concentrations (and nitrite concentrations on a subset
of samples) were measured on a Lachat Quick Chem
8000k Flow Injection Analysis system using stand-
ard methods (Parsons et al., 1984). Underway nutrient
samples were measured every 1.5 min. Ammonium
concentrations were not determined. Chlorophyll a
(Chl a) measurements were performed at sea every 30
to 60 min using fluorometry following the methods
outlined in Parsons et al. (1984) and Welschmeyer
(1994). Chl a size fractionation studies were carried
out by gentle filtration using 25-mm diameter 8-Am
Poreticsk PCTE filters and Whatman GF/F filters.
Filters were extracted in 90% acetone at 20 8C for
12 h and measured on a Turner 10 AU fluorometer
using the nonacidification method (Welschmeyer,
1994). F v/F m measurements were made on the surface
transects and the bottle CTD casts during daylight
hours with a Chelsea Instruments system.
3. Results
3.1. Surface transects
The locations of the surface transects and the
bathymetry of the Peru upwelling region are shown in
Fig. 1A and B, respectively. Sampling during surface
transects occurred approximately from 0800 to 2300 h
local time. During some nights, the ship continued to
steam without sampling, thus resulting in breaks in the
data sets (e.g., between transects 6, 7, and 8). For
transects near the Peru coast, CTD profiles and
sampling on station were carried out during the night
and the ship did not advance again until the next
morning, thus minimizing any geographical breaks in
data.
3.1.1. Transects 6, 7, 8, 9, and 10 (Fig. 2)
Transect 6 began in the equatorial Pacific (2.88S,
86.88W) southeast of the Galapagos Islands in
relatively warm (20 8C) and lower salinity (34.9)
HNLC water of the South Equatorial Current (SEC;
Fig. 2A and B). The nitrate concentration was
approximately 8 AM, silicic acid 4 AM, and phosphate
0.7 AM (Fig. 2C, D, and H). Dissolved Fe averaged
0.1 nM and Chl a concentrations were between 1 and
2 Ag/l (Fig. 2E and G). Type-4 haptophytes (e.g.,
Phaeocystis globosa, Wright et al., 1996) dominated
the phytoplankton community in these waters (DiTul-
lio et al., submitted for publication). During transect 6,
the ship moved into cooler (19 8C) higher-salinity
(35.1) water characteristic of the offshore Peru
Current (the Peru Current is often referred to as the
Humboldt Current or the Peru/Chile Current). Nitrate
concentrations increased to 13 AM, silicic acid to 7
AM, and phosphate to 1.1 AM. Transect 7 continued
through the core of the Peru Current. Nitrate concen-
trations increased to 15 AM, silicic acid to 8 AM, and
phosphate to 1.2 AM. Transect 7 appeared to be in
classic HNLC waters, with dissolved Fe averaging
only 0.05 nM and Chl a averaging only 0.5 Ag/l. The
phytoplankton community on Transect 7 was domi-
nated by Prochlorococcus (DiTullio et al., submitted
for publication).
Transect 8 came within 100 km of the coast of
northern Peru and within 50 km of the shelf break (Fig.
1A and B). At approximately 82.88W longitude, there
was a decrease in temperature and salinity indicative of
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103 87

Fig. 2. Data from surface transects 6 to 10: temperature (A), salinity (B), nitrate (C), silicic acid (D), dissolved Fe (E), depth (F), chlorophyll a
(G) and phosphate (H).

a filament of upwelled water that had been advected
offshore. Dissolved Fe increased slightly to 0.1 nM and
Chl a markedly increased to values of 4 to 6 Ag/l. Along
with this increase in Chl a, there was a tripling of the
average geometric size of the large eukaryotic fraction
of the phytoplankton biomass (Trick, personal com-
munication) and diatoms dominated the phytoplankton
community (DiTullio et al., submitted for publication).
There was a corresponding drop of roughly 5 AM
nitrate and silicic acid. Interestingly, there was also a
decrease in phosphate concentration, but it was a
disproportionately high decrease with respect to Red-
field ratio stoichiometry. Within a few tenths of a
degree longitude, temperature and salinity increased
again, Chl a and dissolved Fe dropped back down, and
nitrate, silicic acid, and phosphate increased (Fig. 2).
88 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
During transect 9, the shelf break was encountered
at 81.18W longitude (Fig. 2F). Coincident with the
sharp shelf break at 90 m depth, dissolved Fe
increased markedly in the surface waters and became
highly variable over the shelf region, nitrate, and
silicic acid decreased abruptly and Chl a increased
and reached values of 10 Ag/l. The phytoplankton
community in these high Chl a waters near the shelf
break consisted of 80–90% diatoms (DiTullio et al.,
submitted for publication).
Transect 10 continued over the inner shelf region
and to within 2 km of the shoreline where cold,
freshly upwelled waters were encountered. At the
near-shore portion of Transect 10, the temperature
dropped to 15 8C, salinity was close to 35.0, nitrate
and silicic acid were both approximately 30 AM,
phosphate increased to 2.3 AM, and dissolved Fe was
approximately 10 nM. In these freshly upwelled
surface waters, dissolved oxygen was only 20% of
saturation, and Chl a only 0.5 Ag/l. All of these
parameters are indicative of extremely fresh upwelled
water, sampled prior to significant phytoplankton
bloom development. Photosynthetic competency in
these freshly upwelled near-shore waters was rela-
tively low with F v/F m values of 0.2 to 0.3 (DiTullio et
al., submitted for publication). Although Fe and
macronutrient concentrations were nonlimiting, there
appeared to be a short lag period for phytoplankton to
acclimate to the higher light levels experienced upon
upwelling and to bshift upQ nutrient assimilation rates
(MacIsaac et al., 1985). Moving back offshore along
transect 10, Chl a increased to 10 Ag/l, and there were
corresponding decreases in nitrate, silicic acid,
phosphate, and dissolved Fe associated with the
diatom bloom. F v/F m values increased to 0.5 in the
diatom-dominated region of Transect 10 where
relatively high Chl a concentrations (ca. 10 Ag/l)
were observed.
3.1.2. Transects 11 and 12 (Fig. 3)
Transects 11 and 12 comprise a second surface
transect across the broad region of the northern Peru
shelf beginning near shore and extending offshore
into blue water (Fig. 1A and B). Transect 11 was
initiated within 2 km offshore in waters with surface
temperatures as low as 15.2 8C, salinity below 35.0,
nitrate 23 AM, silicic acid 27 AM and phosphate 2.3
AM (Fig. 3A, B, C, D, and G). The Chl a
concentration was only 2 Ag/l (Fig. 3F) and dissolved
oxygen was approximately 50% of saturation in this
near-shore upwelled water. Dissolved Fe varied from
8 to 16 nM in the near-shore upwelled waters (Fig.
3E). Moving offshore, there was a rapid rise in
temperature, a rapid decrease in nitrate, silicic acid,
phosphate and dissolved Fe, and a massive phyto-
plankton bloom with Chl a concentrations up to 27
Ag/l. The bloom was N95% diatoms (based upon
pigment data), with F v/F m ratios exceeding 0.6
(DiTullio et al., submitted for publication). This
region is a prime example of the bbrown waters of
PeruQ described by Strickland et al. (1969). In the
surface waters with the highest Chl a concentrations,
silicic acid was drawn down to less than 1 AM and
phosphate to 0.2 AM. There was a 1:1 correspondence
between the Chl a increase (in Ag/l) and the silicic
acid drawdown (in AM). The 14 to 15 AM decrease in
nitrate (plus nitrite) was half the decrease in silicic
acid and only half the decrease expected relative to
the 2 AM decrease in phosphate.
Moving further offshore (at 79.58W longitude),
Chl a dropped to levels of 4 to 5 Ag/l and dissolved Fe
was on the order of 0.5 nM, while nitrate concen-
trations increased to 15–18 AM, silicic acid increased
to approximately 10 AM and phosphate to 1.3 AM
(Fig. 3). Cryptophytes dominated the phytoplankton
assemblage in these waters between 79.58W and
80.08W (DiTullio et al., submitted for publication).
At the end of Transect 11 and the beginning of
Transect 12 (about 80.48W), there was a salinity and
temperature increase indicative of a warm-core eddy.
The dissolved Fe increased to ~1 nM in the core of
this higher salinity feature.
Further offshore in Transect 12 was a relatively
homogeneous region of the Peru Current with
elevated macronutrients (nitrate 16 AM, silicic acid 8
to 9 AM and phosphate 1.2 AM), low dissolved Fe (0.1
nM) and low Chl a (1 Ag/l). This was an HNLC, bblue
waterQ region of the Peru Current. The dominant
phytoplankton was the Type-4 haptophyte, Phaeocys-
tis globosa (DiTullio et al., submitted for publication).
This outer part of Transect 12 was another HNLC area
in which Hutchins et al. (2002) demonstrated Fe
limitation in shipboard incubation experiments.
At the extreme outer part of Transect 12, the
temperature and salinity increased markedly as the
northeastern boundary of the South Pacific central
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103 89

Fig. 3. Data from surface transects 11 and 12: temperature (A), salinity (B), nitrate (C), silicic acid (D), dissolved Fe (E), chlorophyll a (F) and
phosphate (G). The dissolved iron was determined by two separate methods: (o) adsorptive cathodic stripping voltammetry and (w) solvent
extraction/graphite furnace atomic absorption. See text for details.

gyre was encountered. In this case, macronutrients
remained elevated, dissolved Fe remained extremely
low, and HNLC conditions continued into the gyre
waters. Thomas (1979) has shown that these HNLC
conditions continue offshore into lower-latitude
regions of the South Pacific gyre for over 1000 km.
3.1.3. Transect 13 and 14 (Fig. 4)
Transects 13 and 14 comprised a surface transect
off central Peru over another region with a relatively
broad continental shelf (Fig. 1A and B). Transect 13
originated near shore in waters with relatively low
salinity (35.0), but with elevated temperatures (16 to
90 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103

17 8C) compared to further offshore (Fig. 4A and B).
These near-shore waters appeared to have warmed
and aged, allowing the nitrate and silicic acid to be
depleted and the Chl a to increase to about 20 Ag/l
(Fig. 4C, D, and F). Suboxic (e.g., b5 AM dissolved
oxygen and evidence of denitrification), subsurface
waters were within 22 m of the surface. The difference
between Transect 13 and 11 is that the near-shore
waters of Transect 13 had warmed and a phytoplank-
ton bloom had already occurred. Moving offshore, the

Fig. 4. Data from surface transects 13 and 14: temperature (A), salinity (B), nitrate (C), silicic acid (D), dissolved Fe (E), chlorophyll a (F) and
phosphate (G). The dissolved iron was determined by two separate methods: (o) adsorptive cathodic stripping voltammetry and (w) solvent
extraction/graphite furnace atomic absorption. See text for details.
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
temperature dropped to between 15 and 16 8C, and the
macronutrients and Chl a were highly variable.
Extremely high Chl a concentrations (20 to 45 Ag/l)
were observed at longitude 76.68W to 76.98W and the
biomass was dominated by large diatoms (Trick,
personal communication). Based upon pigment anal-
ysis, diatoms composed N90% of the phytoplankton
biomass in this region and F v/F m ratios of ~0.6 were
observed (DiTullio et al., submitted for publication).
Surface waters with Chl a values this high provide an
excellent example of the bbrown waters of PeruQ—
visually, the color of the sea was a chocolate brown.
In the surface waters with Chl a concentrations above
35 Ag/l, the nitrate was drawn down to concentrations
approaching 1 AM. In this brown water region with
such an extensive diatom bloom, the dissolved Fe had
also been depleted to relatively low concentrations as
a result of the high biomass accumulation and
resultant high iron assimilation.
In this high-Chl brown water region between
76.68W and 77.08W, we observed a difference in the
dissolved Fe concentrations determined by the two
analytical methods (Fig. 4E). The shipboard ACSV
method determined dissolved Fe on samples that had
been acidified and microwaved prior to analysis,
while the dithiocarbamate/solvent extraction GFAAS
method determined dissolved Fe on samples that had
been stored acidified in excess of 6 months. In other
water types, we observed excellent agreement
between these two methods. However, in these
extremely high-Chl a brown waters, the ACSV
method determined approximately 0.1 nM dissolved
Fe, while the GFAAS method determined approx-
imately 0.8 nM dissolved Fe. In these extremely high-
chlorophyll brown waters, the ACSV method had an
interfering peak appear in the iron region of the
voltammograms. The Fe(SA)2 reduction peak was on
the shoulder of this interfering peak and a marked
decrease in sensitivity was observed. It appeared that a
high concentration of surfactant or interferent was
present in these intense bloom conditions. This
discrepancy between results obtained by the two
methods and this difference was only obvious in this
extreme, high-Chl, brown water region of Transect 13.
Moving offshore past longitude 77.28W in Transect
13, the temperature and salinity both increased
markedly as we moved into a new water mass and
left the brown water behind. This offshore water of the
91
Peru Current was low in Chl a and contained nitrate at
concentrations between 10 and 15 AM. By 788W, we
had moved further into the offshore region of the Peru
Current and the nitrate and silicic acid were both
about 5 AM and the phosphate 1 AM. The dissolved
Fe was a few tenths of a nanomolar in concentration.
The Chl a concentration was 1 Ag/l or less. We
appeared to be in another blue water HNLC water
mass, although this time, the offshore dissolved Fe
concentrations (~0.3 nM) were slightly higher than
observed on the other transects.
3.1.4. Transect 15 and 16 (Figs. 5 and 6)
Transects 15 and 16 took place off southern Peru in
a region with a narrow shelf (Fig. 1A and B). Transect
15 started near shore in upwelled surface waters with
a temperature of ~14 8C and salinity between 34.70
and 34.85 (Fig. 5A and B). These were recently
upwelled waters with nitrate and silicic acid between
15 and 20 AM, dissolved Fe between 1 and 2 nM, and
Chl a between 1 and 1.5 Ag/l (Fig. 5C, D, E, and F).
As we moved offshore in Transect 15, the temperature
and salinity both rose markedly as the offshore Peru
Current was entered. Nitrate and silicic acid concen-
trations dropped to about 5 AM in this offshore water
mass. In an attempt to find surface waters with higher
macronutrients and low dissolved Fe, we changed
course and went parallel to the coast and eventually
reentered colder, lower-salinity, macronutrient-rich
surface water.
Transect 16 returned parallel to the coast in a
southeast direction (Fig. 1A) and entered in and out of
colder, lower-salinity, nutrient-rich waters (Fig. 6A, B,
C, and D). Temperature and salinity covaried remark-
ably closely during both Transects 15 and 16. During
Transect 16, Chl a varied between 0.4 and 1.0 Ag/l (Fig.
6F) and dissolved Fe was generally between 0.1 and 0.2
nM (Fig. 6E). During Transect 16, we sampled freshly
upwelled waters with low dissolved Fe concentrations
(0.1 to 0.2 nM) that would result in Fe-limited
conditions for diatom blooms almost from the start.
3.2. Vertical profiles/subsurface waters
In an attempt to characterize subsurface source
waters prior to upwelling in the Peru region, we
obtained a series of vertical profiles. Fig. 7A presents
dissolved Fe data from three vertical profiles collected
92 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103

Fig. 5. Data from surface transect 15: temperature (A), salinity (B), nitrate (C), silicic acid (D), dissolved Fe (E), chlorophyll a (F), and
phosphate (G).

along Transects 11 and 12 off northern Peru. The first
vertical profile was taken over the broad inner shelf
off northern Peru, with a bottom depth of 58 m
(Station 18, Fig. 7B). At this station, near-bottom
samples from 52 m depth in the middle of the suboxic
waters had a dissolved Fe concentration of 48 to 51
nM. Fig. 8E and F presents the temperature, percent
(%) oxygen saturation, nitrate, nitrite, and phosphate
profiles from this station. The waters were suboxic at
depths deeper than 28 m, and values of 8 AM nitrite
were observed in the suboxic zone. There was a
gradient of dissolved Fe from 10 to 50 nM at depths
of 24 to 52 m, with concentrations increasing towards
the sediments within the suboxic waters (Fig. 7A).
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103 93

Fig. 6. Data from surface transect 16: temperature (A), salinity (B), nitrate (C), silicic acid (D), dissolved Fe (E), chlorophyll a (F), and
phosphate (G).

A second profile was obtained about 200 km
offshore just off the shelf break over the Peru Trench
(Station 19, Fig. 7B). At this station, the surface
mixed layer extended down to 40 m depth and the
waters were suboxic at depths from 70 to 130 m with
nitrite concentrations of 7 AM (Fig. 8C and D). Here,
the dissolved Fe at 80 m depth in the core of the
suboxic waters was 2.3 to 3.8 nM (Fig. 7A). A third
profile was obtained at approximately 400 km off-
shore at the outer edge of the Peru Current at the
boundary of the South Pacific central gyre (Station 20,
Fig. 7B). Here, the mixed layer extended to a depth of
50 m and oxygen-depleted suboxic waters were
observed below 75 m and extending to 200 m (Fig.
94 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Fig. 7. Vertical profiles of dissolved Fe at three stations (18, 19, and 20) along transect 11 and 12: the profiles include an inner shelf station (18),
a station just off the shelf break (19), and a station in the offshore Peru Current (20). The station locations are shown in the figure on the right.
Also shown is the location of a station over the shelf off central Peru (25).
8A). Nitrite concentrations were 6 AM in the core of
the suboxic zone (Fig. 8B). In this offshore profile,
dissolved Fe concentrations were less than 1 nM (Fig.
7A). The subsurface water at all three stations along
this transect was suboxic with between 6 and 8 AM
nitrite and with nitrate anomalies clearly evident. At
each station, a strong nitrate anomaly was observed,
but there was always at least 10 AM nitrate remaining
in the suboxic waters. The existence of at least 10 AM
nitrate, together with the 6 to 8 AM nitrite indicates
that suboxic conditions, not anoxic conditions, existed
(Rue et al., 1997). The high dissolved Fe concen-
trations observed near shore in proximity with the
shelf sediments do not extend offshore into the
ocean’s interior although high nitrite and suboxic
conditions still persisted in these offshore subsurface
waters.
A subsurface near-bottom sample was also
obtained over a wide shelf region off central Peru.
We chose a site with a bottom depth of 65 m and
obtained a sample from a depth of 60 m (Station 25,
Fig. 7B). At this station, the waters were suboxic
beneath a depth of 28 m with elevated nitrite present
along with 10 AM nitrate (Fig. 8G and H). The
dissolved Fe at this central Peru location was 118 nM.
This is the highest concentration of dissolved Fe we
have observed in the coastal waters of any eastern
boundary region. This was also the apparent source
water for the remarkable brown water bloom con-
ditions (with Chl a values up to 45 Ag/l) observed in
the overlying near-shore waters of Transect 13.
The region off southern Peru has a narrow
continental shelf (Fig. 1B). The subsurface waters
are still suboxic, but do not have a broad expanse of
shelf sediments with which they are in contact. Near-
bottom samples (within 5 to 10 m of the bottom) were
collected from depths of 40 to 65 m on transects T-16,
T-17 and T-18. In each case, the surface waters had
low dissolved Fe—less than 0.2 nM. Near-bottom
dissolved Fe concentrations in this region with a
narrow continental shelf ranged from 1.4 to 4.3 nM.
The highest concentration observed in the near-
bottom shelf waters in this southern region were more
than an order-of-magnitude lower than the concen-
trations observed over the broad shelf region off
northern and central Peru.
3.3. Size fractionated Chl a data
Size-fractionated Chl a data were obtained on a
subset of the Chl a determinations and are presented
in Fig. 9. When the Chl a concentration was greater
than 5 Ag/l, the percentage of chlorophyll in the N8
Am size fraction was N75% and averaged close to
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Fig. 8. Vertical profiles of temperature (o) and percent (%) oxygen saturation (w) for the offshore Peru Current station 20 (A), station 19 just off the shelf break (C), and stations 18
and 25 located over the inner shelf (E and G, respectively). Nitrate (.), Nitrite ( S) and phosphate (o) are presented in B, D, F and H, corresponding to A, C, E and G, respectively.

95
96 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Fig. 9. Total chlorophyll a versus the percent chlorophyll a in the
N8 um size fraction for the Peru upwelling system.
90%. These high Chl a samples were dominated by
large diatoms (Trick, personal communication; DiTul-
lio et al., submitted for publication). The low-Chl
bblue waterQ regions of the Peru upwelling system
with Chl a concentrations less than 2 Ag/l were
dominated by the b8 Am size fraction with large
diatoms being rarer and the phytoplankton community
being composed chiefly of cryptophytes, prasino-
phytes and haptophytes (DiTullio et al., submitted for
publication).
4. Discussion
4.1. bBlue and brown watersQ of the Peru Current
4.1.1. bBlue watersQ
The offshore regions of the Peru/Chile Current
were characterized by temperatures between 17 and
20 8C, salinity between 35.0 and 35.2, nitrate
concentrations of approximately 15 AM, silicic acid
concentrations between 5 and 10 AM, phosphate
concentrations between 1 and 1.5 AM, and Chl a
concentrations less than 2 Ag/l. This included regions
of transects 6, 7 and 8 from longitudes of 868W to
838W, transect 12 from 81.58W to 808W, and Trans-
ects 14 and 15 from 798W to 77.58W. These offshore
regions of the Peru/Chile Current appeared to be iron-
limited, high-nitrate and lower than expected chlor-
ophyll (HNLC) regimes.
Transect 7 (Figs. 1A and 2) and transect 12 (Figs.
1A and 3) provide typical examples of HNLC bblue
watersQ of the Peru Current. These transects are
located off the shelf between 100 and 500 km from
the coast of northern and central Peru and are
examples of the offshore waters of the Peru upwelling
region that Thomas (1979) found bremarkableQ in that
the high nutrient concentrations observed in these
surface waters had not been depleted by phytoplank-
ton assimilation. Transect 7 was relatively homoge-
neous with nitrate concentrations averaging 15 AM,
silicic acid 8 AM, and phosphate 1.2 AM. Chl a
concentrations were only 0.5 Ag/l and dissolved Fe
ranged from 0.05 to 0.1 nM. Transect 12 was also
relatively homogeneous with elevated macronutrients
(nitrate 16 AM, silicic acid 8 to 9 AM, and phosphate
1.2 AM), low dissolved Fe (0.05 to 0.1 nM), and
relatively low Chl a (1 Ag/l).
Both of these HNLC bblue waterQ regions were
sampled and shown to be Fe-limited regimes based on
shipboard Fe-enrichment grow-out experiments
(Hutchins et al., 2002). In those experiments, Chl a,
F v/F m, and the drawdown of nitrate and phosphate
increased in direct proportion to the amount of Fe
added. In the Fe addition experiment from the
Transect 7 region (referred to as the Humboldt Current
Station in Hutchins et al., 2002), there was a
pronounced Fe-mediated shift towards diatoms in the
Fe addition incubations. Total diatom numbers were
dominated by small pennates, however, larger species
such as the large unicellular centric diatom (Cosci-
nodiscaceae) increased dramatically with increasing
Fe availability and contributed a substantial share of
the increase in total biomass in the grow-outs with Fe
additions. In the grow-out experiments carried out in
the offshore area of Transect 12 (a region referred to
as the Peru Current Station in Hutchins et al., 2002),
there was also an Fe-mediated shift towards diatoms
with the smaller pennates being dominant both
numerically and in biomass. At this station, P.
globosa flagellated cell numbers also increased in
the Fe addition experiments.
DiTullio et al. (submitted for publication) report
F v/F m ratios from these various transects in the Peru
upwelling regime as a measure of photosynthetic
competency. These ratios have been shown to be a
rapid and sensitive indicator of iron stress in HNLC
regions (Behrenfeld et al., 1996). As an example of
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103 97

how the F v/F m ratio varies in the Peru upwelling
regime, we present a small subset of results from
DiTullio et al. (submitted for publication) for two
portions of one of the key surface transects (Transect
11/12) that traversed both brown and blue water
regions. A portion of Transect 12 (Fig. 10E, F, G, and
H) provides a good example of a low-iron bblue
waterQ region of the Peru Current. In this region, F v/
F m ratios ranged from values of 0.15 to 0.30. These
low F v/F m values are indicative of resource limitation
of the phytoplankton community—suggestive of Fe-
limited conditions (Behrenfeld et al., 1996; Behren-
feld and Falkowski, 1999). Indeed, the dissolved Fe
concentrations were between 0.05 and 0.1 nM—

Fig. 10. Selected subsets of data for Transect 11 and Transect 12 representing bblue waterQ and bbrown waterQ regions of the Peru upwelling
system. Presented are underway fluorescence and chlorophyll a (A and E), F v/F m (B and F), nitrate (o) and silicic acid ( R) (C and G), and
dissolved Fe (D and H).
98 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
concentrations that would be limiting to populations
of coastal diatoms (Bruland et al., 2001; Firme et al.,
2003). Macronutrient concentrations in this region
were nonlimiting.
DiTullio et al. (submitted for publication) also
report flavodoxin and ferredoxin values determined
on phytoplankton samples collected during these
various transects. A response common to phytoplank-
ton subjected to Fe-deplete conditions is the sub-
stitution of ferredoxin with flavodoxin, a non-Fe
containing functional equivalent (LaRoche et al.,
1996). Elevated expression of flavodoxin in a range
of diatom species tested is induced by Fe limitation,
suggesting that ferredoxin and flavodoxin can be used
as diagnostic markers of physiological Fe-status in the
ocean. In the blue water region of Transect 12, only
flavodoxin was observed in algal populations domi-
nated by haptophytes indicating Fe-stress conditions
(DiTullio et al., submitted for publication).
Hutchins et al. (2002) performed a series of Fe-
amendment experiments at a station at the outermost
part of Transect 12 in the low Fe, HNLC, blue water
region (referred to as the Peru Current Station). Initial
F v/F m ratios in their samples were 0.27 and in the
bottles to which Fe was added, rapid increases in F v/
F m were observed, peaking on day 3. Bottles with an
addition of 2.5 nM dissolved Fe reached maximum
F v/F m ratios of 0.47. These increases in F v/F m are
consistent with Fe being the limiting nutrient and with
the lack of iron inhibiting the photosynthetic com-
petence or F v/F m of the phytoplankton community in
this region. The +2.5 nM dissolved Fe addition
resulted in Chl a increasing from just under 1 Ag/l
to a maximum of 8 Ag/l on day 4 of the incubation.
The increase in Chl a was primarily due to pennate
diatoms and haptophytes that responded to the Fe
additions at this offshore station (Hutchins et al.,
2002).
4.1.2. bBrown watersQ
The near-shore region of Transect 11 provides a
good example of high-Chl bbrown waters of Peru.Q In
the high-Chl a, high-iron regions, F v/F m ratios were
greater than or equal to 0.6 (Fig. 10A, B, C and D).
These are remarkably high F v/F m ratios for natural
populations and indicate exceptionally high photo-
synthetic competency, suggesting that the phytoplank-
ton community was growing at optimal rates. In this
case, there was adequate iron available to drive the
macronutrient assimilation, resulting in a large diatom
bloom in these shelf waters. Ferredoxin, rather than
flavodoxin, was dominant in phytoplankton from this
region (DiTullio et al., submitted for publication).
Thus, both of these biological indicators point to Fe-
replete conditions in this bbrown waterQ part of
Transect 11. This massive diatom bloom is consistent
with the time and space domain reported by MacIsaac
et al. (1985) who state that the cycle from initial
upwelling to bloom development and nutrient draw-
down in this region is completed in 8 to 10 days
within 30 to 60 km off the coast.
Codispoti et al. (1982), in a comparison of eastern
boundary upwelling regions, concluded that the high
biological productivity of the Peruvian upwelling
system may be linked to the ability of the phyto-
plankton to take up and utilize nitrate at an extra-
ordinary rate. We suggest that a critical factor leading
to these especially high rates off Peru is the co-
occurrence of high macronutrient and high iron
concentrations in the euphotic zone. This suggestion
is supported by the maximal F v/F m ratios in excess of
0.6. This situation occurs over the broad shelf regions
off northern and central Peru where high-chlorophyll
brown waters are found.
4.2. Sources of dissolved Fe in the Peru upwelling
system
Extremely high concentrations of dissolved Fe
were observed in subsurface suboxic waters in contact
with shelf sediments off northern and central Peru at
depths of 40 to 100 m. These near-bottom waters over
the broad shelf areas off northern and central Peru had
dissolved Fe concentrations ranging from 48 to 118
nM. These values are consistent with those observed
by Hong and Kester (1986) who reported concen-
trations of dissolved Fe(II) greater than 50 nM in the
suboxic near-bottom shelf water at depths of 40 to 100
m within 5 to 10 km of the coast off central Peru.
Although the suboxic conditions and high nitrite
values persist far offshore (Fig. 7C and D), the
dissolved Fe concentrations rapidly diminish away
from the shelf (Fig. 6). This is consistent with the
redox conditions of Fe in suboxic waters (Rue et al.,
1997), in that the tendency is for particle-reactive
Fe(III) to be the thermodynamically stable form of
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
iron in the suboxic waters of the eastern tropical
Pacific. It takes anoxic conditions for Fe(II) to be the
thermodynamically stable form of iron (Rue et al.,
1997). Hong and Kester (1986) argued for a shelf
sediment source for the high concentrations of Fe(II)
they observed in the near-bottom shelf waters. It
appears that the organic-rich shelf sediments can
provide a substantial source of iron to the near-shore
suboxic water column that can be upwelled into the
surface photic zone.
Both analytical techniques used in this study for
the determination of dissolved Fe measure the sum of
Fe(II) and Fe(III) and do not distinguish between
dissolved iron in these two oxidation states. We did
examine the electrochemical lability of the dissolved
Fe in these subsurface waters. At the shelf station off
northern Peru, we observed the following results: at
52 m depth, 39 nM of the 48 nM total dissolved Fe
was labile, and at 40 m depth, 16 of the 21.5 nM
dissolved Fe was labile. At the shelf station off central
Peru, 73 nM was labile out of 118 nM total dissolved
Fe. Labile dissolved iron is that iron that exchanges
with the competing ligand SA (27.5 AM) when there
has been no pretreatment of the sample (e.g., no
acidification or microwave pretreatment). This large
fraction of labile Fe is consistent with the bulk of the
dissolved Fe in this suboxic shelf water existing as
soluble Fe(II), rather than as a less-reactive Fe(III)
colloidal form.
The near-bottom waters in the narrow shelf region
off southern Peru, although they were also suboxic,
contained relatively low concentrations of dissolved
Fe—values ranging from 1.4 to 4.6 nM. As a result,
only waters upwelling over the broad shelf regions off
northern and central Peru are likely to have high
concentrations of the micronutrient Fe entrained along
with high concentrations of macronutrients. Transect
10 (Fig. 2E) had dissolved Fe concentrations of 10 nM
in freshly upwelled waters. Transect 11 (Fig. 3E)
obtained over northern Peru had dissolved Fe con-
centrations of 6 to 16 nM in the high-nutrient freshly
upwelled waters. In Transect 13, a massive phyto-
plankton bloom with Chl a concentrations up to 45
Ag/l had already occurred along the coast by the time
we sampled (Fig. 4F). When this area was sampled,
the winds had previously died down and the system
was in the midst of a relaxation event, with a massive
diatom bloom already present. This extensive phyto-
99
plankton bloom would have drawn down the dis-
solved Fe originally present in freshly upwelled waters
of this region, and concentrations of only 1 to 3 nM
were observed at this time. In Transect 15 off southern
Peru (Fig. 5E), freshly upwelled waters contained 1 to
2 nM dissolved Fe, while in Transect 16 (Fig. 6E),
freshly upwelled waters had only 0.1 to 0.3 nM
dissolved Fe and diatoms comprised b20% of the total
Chl a biomass. Thus, upwelled waters off Peru vary
markedly in the amount of dissolved Fe initially
entrained along with the macronutrients. High rates of
diatom production on the Peruvian shelf are dependent
upon an adequate supply of dissolved Fe. As a result
of the broad range of initial conditions occurring in the
coastal upwelling waters off Peru, a mosaic of Fe-
replete to Fe-deplete conditions develops, similar to
what has been observed off central California (Hutch-
ins et al., 1998; Bruland et al., 2001).
Bruland et al. (2001) discussed how phytoplankton
Fe assimilation varies in a non-Redfield manner
depending upon the availability of iron. They exam-
ined laboratory data for coastal diatoms generated by
Sunda and Huntsman (1995) and estimated Fe/C
ratios (Amol Fe/mol C) as a function of available
[FeV]. This variable relationship can be used to
evaluate the drawdown of iron relative to nitrate.
Fig. 11A presents scenarios of dissolved iron and
nitrate drawdown for various starting conditions
where an upwelled parcel of water undergoes a
phytoplankton bloom. The nitrate is plotted as a
function of [FeV], where in this case, [FeV] includes all
the biologically available iron. In the defined labo-
ratory studies of Sunda, the [FeV] referred to the sum
of inorganic forms of Fe(III) ([Fe(OH)2+], [Fe(OH)3o]
and [Fe(OH)4 ]), with Fe chelated to EDTA being
nonavailable. The scenario starting with 12 nM Fe and
30 AM nitrate represents initial conditions that might
be representative of upwelled waters in the near-shore
regions off northern and central Peru. In this case, the
initial Fe/C ratio assimilated by diatoms would be
approximately 300 Amol/mol, and Fe is rapidly drawn
down relative to nitrate. As the Fe concentration
decreases, the Fe/C assimilation ratio of the diatoms
markedly decreases, and the slope of Fe to nitrate
removal decreases. As the Fe drops below 1 nM, the
Fe/C ratio drops to 20 Amol/mol. Eventually, as the Fe
becomes depleted to the point where it limits the
growth rate, the Fe/C assimilation ratio of coastal
100 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Fig. 11. The drawdown of dissolved Fe relative to nitrate for
different starting concentrations (A) for the situation where all the
dissolved iron is readily available. This was prepared using the
relationship between Fe/C and Fe/N assimilation and the dissolved
FeV concentration, [FeV], by coastal diatoms observed in laboratory
studies by Sunda and Huntsman (1995) and discussed in Bruland et
al. (2001). Also shown (B) are scenarios where only 25% of the
dissolved iron is readily available and the effect of remineralization
of phytoplankton that had originally assimilated iron under luxury
uptake conditions.
diatoms can drop to approximately 10 Amol/mol or
less. The scenario of 30 AM nitrate and 2 nM [FeV] is
representative of conditions observed in the southern
Peru upwelling regime. Here, iron rapidly becomes
the limiting nutrient.
The net result of this variable, non-Redfield
removal of Fe relative to nitrate is that what appears
to be a rich source of dissolved Fe to complement the
macronutrients can be rapidly assimilated, leaving
behind waters that are still relatively high in macro-
nutrients, but Liebig-limited by Fe. As the bloom
develops and nutrients are assimilated, the system
evolves towards Fe limitation where larger diatoms
are unable to grow at high rates because of their high-
iron half-saturation constants for optimal growth
relative to the low Fe concentrations (Sunda and
Huntsman, 1995). This preferential depletion of iron
in coastal waters can result in the bblue waters of
PeruQ or high nitrate but lower than expected
chlorophyll areas within the Peru upwelling regime.
Two other scenarios are also shown in Fig. 11B.
The first is what would occur if only 25% of the
dissolved Fe was biologically available to the diatom
community, with 75% being in some nonavailable
form. This is modelled so that the nonavailable form
is in equilibrium with the available form and main-
tains the available iron at 25% of the total dissolved
iron. In this case, with a lower concentration of iron
being available, the diatoms assimilate iron at a lower
Fe/C ratio and the 12 nM initial dissolved Fe is
adequate to allow complete drawdown of the 30-AM
nitrate. Not shown in Fig. 11B is what could happen if
luxury uptake of Fe by diatoms during Fe-replete
conditions (e.g., assimilation of Fe at an Fe/C ratio of
100 to 300 Amol/mol) allowed the diatoms to sustain
additional high growth rates for multiple generations
during Fe-deplete conditions. For example, iron stored
as ferredoxin or other Fe-rich compounds during
initial phases of luxury uptake could be passed on to
the progeny and become available to fulfill the Fe
requirements of future generations of diatoms. Also
shown in Fig. 11B is an attempt to take into account
remineralization of phytoplankton biomass by grazing
during mid to later stages of the bloom, that would
regenerate Fe and N with higher Fe/C ratios (and Fe/
N) assimilated earlier in the bloom (Hutchins et al.,
1993; Hutchins and Bruland, 1994). For example,
high (~200 Amol/mol) Fe/C assimilation ratios in
diatoms grown during the early stages of the bloom
can, upon regeneration, fuel a proportionally higher
diatom biomass assimilating Fe at Fe/C ratios of 20
Amol/mol under conditions of low Fe availability.
Both of these scenarios involving initial luxury uptake
would allow additional drawdown of nitrate for a
given initial Fe concentration during the latter stages
of a bloom.
In addition, if the initial upwelled water continued
to be bfedQ or mixed with an additional source of
upwelled water that was rich in Fe (over the shelf),
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
this could keep providing a source of water with high
Fe relative to nitrate that could keep diatoms growing
at near optimal rates. In fact, a scenario whereby the
available Fe is maintained at approximately 1 to 2 nM
would be more ideal than providing an initial large
concentration of 10 to 20 nM and then shutting off
any further source of nutrients. Although Fig. 11 is a
simplistic model of a diatom bloom developing in an
isolated water parcel of upwelled water, the model still
demonstrates that these coastal upwelling systems
have a strong tendency to be driven towards Fe
limitation, especially as sedimentation of iron-rich
diatom biomass exports Fe out of the surface photic
zone. These processes can lead to the rapid develop-
ment of blue water HNLC conditions.
The massive blooms that occurred in Transect 13
provide an example of non-Redfield removal of
dissolved Fe relative to nitrate. Presumably, high
dissolved Fe concentrations were originally upwelled
to the surface waters along with the macronutrients.
The subsurface waters at this location contained
extremely high concentrations of dissolved iron—in
excess of 100 nM. However, the highest dissolved Fe
concentration observed in surface waters of this
region experiencing relaxation conditions, where a
bloom had already occurred, were only 2 nM. In the
region where the bloom was the most pronounced
(with Chl a concentrations between 35 and 45 Ag/l),
the dissolved Fe was less than 1 nM (Fig. 4E). We
suggest that these upwelled waters had initially
contained extremely high dissolved Fe that had been
subsequently assimilated and depleted by both the
high diatom biomass and luxury Fe-uptake during
early stages of the bloom. Because of the low-wind
relaxation conditions that existed during Transect 13,
we sampled the system just as the bloom was peaking
and iron was depleted.
On every surface transect, once we moved offshore
of the shelf break, the Peru upwelling system became
an iron-limited HNLC or bblue waterQ regime. Any
upwelling or fresh supply of subsurface nutrients
occurring off the shelf break would provide low
concentrations of dissolved Fe along with the
enhanced macronutrients and would quickly evolve
into a Fe-limited HNLC regime.
Modeling studies of dust transport and deposition
(Mahowald et al., 1999) suggest that this offshore
region of Peru extending out into the northeastern
101
portion of the South Pacific subtropical gyre and into
the equatorial Pacific has the lowest dust input of
anywhere in the oceans. Moore et al. (2002) have
developed a model that yields estimates of dissolved
Fe in these waters of only 5 to 30 pM—the lowest
model values predicted for anywhere in the oceans.
This is a large area of the eastern tropical South Pacific
that the model of Moore et al. (2002) predicts should
be severely iron-limited. Several studies (Thomas,
1979; Pena et al., 1994) have shown the persistence of
unused nitrate in surface waters extending far offshore
from the Peru upwelling region, despite irradiance and
stratification conditions favorable to stimulate nutrient
uptake rates. This persistence of unused nitrate results
in an important fraction of the upwelled nitrate being
transported away from the upwelling zone (Fiedler et
al., 1991). The results presented in this paper and the
incubations carried out by Hutchins et al. (2002)
provide field evidence in support of the role of iron
limitation being the key factor leading to the presence
of these large HNLC areas.
MacIsaac et al. (1985) proposed an alternative
explanation of the optimal trace metal balance of the
nutrient-rich water in this region based upon the ideas
of Barber et al. (1971). These authors suggested that
the alongshore flow of the poleward undercurrent
comes into contact with the reducing sediments of the
Peruvian shelf, and that during this contact, the source
water for upwelling gains manganese from the
reduced sediment and loses copper as an insoluble
sulfide. They suggest that increased manganese ion
concentrations and decreased copper ion concentra-
tions provide the trace metal ion balance required for
maximal growth rates by phytoplankton in such
nutrient-rich water. We suggest, however, that the
role of iron as a micronutrient is far more important in
providing the optimum trace metal balance to the Peru
coastal upwelling regime, rather than optimal Mn and
Cu concentrations.
The impetus for this study was to examine the role
of iron limitation in the Peru upwelling regime and the
impact of a broad shelf to serve as an iron trap and
provide an external source of iron in a similar manner
to what we have previously observed off central
California. We have shown unambiguously that iron
plays an important role in the Peru upwelling system.
In northern and central Peru, where a broad shelf
exists, there is a substantial external supply of iron
102 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
from reducing shelf sediments to complement the
macronutrients, and extensive blooms of large dia-
toms lead to the existence of bbrown watersQ. The
southern coastal upwelling regions and offshore
regions of the Peru Current lack this external supply
of iron and the waters evolve into Fe-limited HNLC
bblue watersQ containing relatively low diatom bio-
mass and Chl a.
Acknowledgements
We thank the Captain and crew of the R/V Melville
for their assistance on this research expedition. We
appreciate the support of Jenn Conn for performing
the nutrient measurements, Charlie Trick for his flow
cytometry data, and Laura Lessin for the discrete
Chlorophyll a data. Support was provided by the
National Science Foundation: grants OCE 9811114
and OCE 0137085 to KWB and OCE 9907931 to
GRD.
References
Alheit, J., Bernal, P., 1993. Effects of physical and biological
changes on the biomass yield of the Humboldt Current
ecosystem. In: Sherman, K., Alexander, L.M., Gold, B.D.
(Eds.), Large Marine Ecosystems: V. Stress, Mitigation and
Sustainability. American Association for the Advancement of
Science, Washington, DC, pp. 55 – 68.
Barber, R.T., Smith, R.L., 1981. Coastal upwelling ecosystems. In:
Longhurst, A.R. (Ed.), Analysis of Marine Ecosystems.
Academic Press, New York, pp. 31 – 68.
Barber, R.T., Dugdale, R.C., MacIsaac, J.J., Smith, R., 1971.
Variations in phytoplankton growth associated with the
source and conditioning of upwelling water. Investig. Pesq.
35, 171 – 193.
Bakun, A., 1987. Monthly variability in the ocean habitat off Peru
as deduced from maritime observations, 1953 to 1984. In: Pauly,
D., Tsukayama, I. (Eds.), Peruvian Anchoveta and its Upwelling
Ecosystem, pp. 46 – 74.
Bakun, A., Nelson, C.S., 1991. The seasonal cycle of wind-stressed
curl in subtropical eastern boundary current regimes. J. Phys.
Oceanogr. 21, 1815 – 1834.
Behrenfeld, M.J., Falkowski, P.G., 1999. Widespread Fe limitation
of phytoplankton in the South Pacific Ocean. Science 283,
840 – 843.
Behrenfeld, M.J., Bale, A.J., Kolber, Z.S., Aiken, J., Falkowski,
P.G., 1996. Confirmation of iron limitation of phytoplankton
photosynthesis in the equatorial Pacific Ocean. Nature 383,
508 – 510.
Brink, K.H., Halpern, D., Huyer, A., Smith, R.L., 1983. The
physical environment of the Peruvian upwelling system. Prog.
Oceanogr. 12, 285 – 305.
Bruland, K.W., Rue, E.L., 2001. Chptr 6: Iron: analytical methods
for the determination of concentrations and speciation. In:
Hunter, K.A., Turner, D.R. (Eds.), The Biogeochemistry of Iron
in Seawater. Wiley, pp. 255 – 289.
Bruland, K.W., Franks, R.P., Knauer, G.A., Martin, J.H., 1979.
Sampling and analytical methods for the determination of
copper, cadmium, zinc and nickel in seawater. Anal. Chim.
Acta 105, 233 – 245.
Bruland, K.W., Rue, E.L., Smith, G.J., 2001. Iron and macro-
nutrients in California coastal upwelling regimes: implications
for diatom blooms. Limnol. Oceanogr. 46, 1661 – 1674.
Carr, M.-E., 2002. Estimation of potential productivity in eastern
boundary currents using remote sensing. Deep-Sea Res., Part 2,
Top. Stud. Oceanogr. 49, 59 – 80.
Codispoti, L.A., Dugdale, R.C., Minas, H.J., 1982. A comparison of
the nutrient regimes off northwest Africa, Peru and Baja
California. Reun. Cons. Int. Explor. Mer. 180, 184 – 201.
de Mendiola, B.R., 1981. Seasonal phytoplankton distribution
along the Peruvian coast. In: Richards, F.A. (Ed.), Coastal
Upwelling. American Geophysical Union, Washington, DC,
pp. 348 – 356.
DiTullio, GR., Geesey, M.E., Maucher, J.M., Alm, M.B.,
Riseman, S.F., Bruland, K.W., 2003. Influence of iron on
Algal community composition and physiological status in the
Peru upwelling system. Limnol. Oceanogr. (submitted for
publication).
Eldridge, M.L., Trick, C.G., Alm, M.B., DiTullio, G.R., Rue, E.L.,
Bruland, K.W., Hutchins, D.A., Wilhelm, S.W., 2004. The
response of the marine phytoplankton community to a manip-
ulation of bioavailable iron in HNLC waters of the subtropical
Pacific Ocean. Aquat. Microb. Ecol. 35, 79 – 91.
Fiedler, P.C., Philbrick, V., Chavez, F.P., 1991. Oceanic upwelling
and productivity in the eastern tropical Pacific. Limnol. Ocea-
nogr. 36, 1834 – 1850.
Firme, G.F., Bruland, K.W., Rue, E.L., Weeks, D.A., Hutchins,
D.A., 2003. Spatial and temporal variability in phytoplankton
iron limitation along the California coast and consequences for
Si, N and C biogeochemistry. Glob. Biogeochem. Cycles 17,
1016.
Fung, I.Y., Meyn, S.K., Tegen, I., Doney, S.C., John, J.G., Bishop,
J.K.B., 2000. Iron supply and demand in the upper ocean. Glob.
Biogeochem. Cycles 14, 281 – 296.
Hebbeln, D., Marchant, M., Freudenthal, T., Wefer, G., 2000.
Surface sediment distribution along the Chilean continental
slope related to upwelling and productivity. Mar. Geol. 164,
119 – 137.
Hong, H., Kester, D.R., 1986. Redox state of iron in the offshore
waters of Peru. Limnol. Oceanogr. 31, 512 – 524.
Hutchins, D.A., Bruland, K.W., 1994. Grazer-mediated regeneration
and assimilation of Fe, Zn, and Mn from planktonic prey. Mar.
Ecol., Prog. Ser. 110, 259 – 269.
Hutchins, D.A., Bruland, K.W., 1998. Iron-limited diatom growth
and Si:N uptake ratios in a coastal upwelling regime. Nature
393, 561 – 564.
 K.W. Bruland et al. / Marine Chemistry 93 (2005) 81–103
Hutchins, D.A., DiTullio, G.R., Bruland, K.W., 1993. Iron and
regenerated production: evidence for biological iron recycling.
Limnol. Oceanogr. 38, 1242 – 1255.
Hutchins, D.A., DiTullio, G.R., Zhang, Y., Bruland, K.W., 1998. An
iron mosaic in the California upwelling regime. Limnol.
Oceanogr. 43, 1037 – 1054.
Hutchins, D.A., Hare, C.E., Weaver, R.S., Zhang, Y., Firme, G.F.,
DiTullio, G.R., Alm, M.B., Riseman, S.F., Maucher, J.M.,
Geesey, M.E., Trick, C.G., Smith, G.J., Rue, E.L., Conn, J.,
Bruland, K.W., 2002. Phytoplankton iron limitation in the
Humboldt Current and Peru. Limnol. Oceanogr. 47, 997 – 1011.
Johnson, K.S., Chavez, F.P., Friederich, G.E., 1999. Continental
shelf sediment as a primary source of iron for coastal
phytoplankton. Nature 398, 697 – 700.
Johnson, K.S., Chavez, F.P., Elrod, V.A., Fitzwater, S.E., Penning-
ton, J.T., Buck, K.R., Waltz, P.M., 2001. The annual cycle of
iron and the biological response in central California coastal
waters. Geophys. Res. Lett. 28, 1247 – 1250.
LaRoche, J., Boyd, P.W., McKay, R.M., Geider, R.J., 1996.
Flavodoxin as an in situ marker for iron stress in phytoplankton.
Nature 382, 802 – 805.
MacIsaac, J.J., Dugdale, R.C., Barber, R.T., Blasco, D., Packard,
T.T., 1985. Primary production cycle in an upwelling center.
Deep-Sea Res. 32, 403 – 529.
Mahowald, N., Kohfeld, K., Hansson, M., Balkanski, Y., Harrison,
S., Prentice, I., Schulz, M., Rodhe, H., 1999. Dust sources and
deposition during the last glacial maximum and current climate:
a comparison of model results with paleodata from ice cores and
marine sediments. J. Geophys. Res. 104, 15895 – 15916.
Martin, J.H., Gordon, R.M., 1988. Northeast Pacific iron distribu-
tions in relation to phytoplankton productivity. Deep-Sea Res.
40, 115 – 134.
Minas, H.J., Minas, M., 1992. Net community production in bhigh
nutrient–low chlorophyllQ waters of the tropical and Antarctic
oceans: grazing vs iron hypothesis. Oceanol. Acta 15, 145 – 162.
Moore, J.K., Doney, S.C., Glover, D.M., Fung, I.Y., 2002. Iron
cycling and nutrient-limitation patterns in surface waters of the
World Ocean. Deep-Sea Res., Part 2, Top. Stud. Oceanogr. 49,
463 – 507.
Nelson, S.S., Neshyba, S., 1979. On the southernmost extension of
the Peru–Chile undercurrent. Deep-Sea Res. 26A, 1387 – 1393.
Parsons, T.R., Maita, Y., Lalli, C.M., 1984. A manual of chemical
and biological methods for seawater analysis. Pergamon.
Pena, M.A., Lewis, M.R., Cullen, J.J., 1994. New production in the
warm waters of the tropical Pacific Ocean. J. Geophys. Res. 99,
14255 – 14268.
Rue, E.L., Bruland, K.W., 1995. Complexation of iron(III) by
natural organic ligands in the central North Pacific as
103
determined by a new competitive ligand equilibration/adsorp-
tive cathodic stripping voltammetry method. Mar. Chem. 50,
117 – 138.
Rue, E.L., Bruland, K.W., 1997. The role of organic complexation
on ambient iron chemistry in the equatorial Pacific Ocean and
the response of a mesoscale iron addition experiment. Limnol.
Oceanogr. 42, 901 – 910.
Rue, E.L., Smith, G.J., Cutter, G.A., Bruland, K.W., 1997. The
response of trace element redox couples to suboxic conditions in
the water column. Deep-Sea Res. 44, 113 – 134.
Scheidegger, K.F., Krissek, L.A., 1981. Zooplankton and nekton:
natural barriers to the seaward transport of suspended terri-
genous particles off Peru. In: Suess, E., Thiede, J. (Eds.),
Coastal Upwelling: Part A. Responses of the sedimentary
regime to present coastal upwelling. Plenum Press, New York,
pp. 303 – 336.
Scholl, D.W., Christensen, M.N., von Huene, R., Marlow, M.S.,
1970. Peru–Chile trench. Sediments and sea-floor spreading.
Geol. Soc. Amer. Bull. 81, 1339 – 1360.
Smetacek, V., 1998. Biological oceanography—diatoms and the
silicate factor. Nature 391, 224 – 225.
Strickland, J.D.H., Eppley, R.W., de Mendiola, B.R., 1969.
Phytoplankton populations, nutrients and photosynthesis in
Peruvian coastal waters. Bol.-Inst. Mar. Peru 2, 4 – 45.
Sunda, W.G., Huntsman, S.A., 1995. Iron uptake and growth
limitation in oceanic and coastal phytoplankton. Mar. Chem. 50,
189 – 206.
Thomas, W.H., 1972. Nutrient inversions in the southeastern
tropical Pacific Ocean. Fish. Bull. 70, 929 – 932.
Thomas, W.H., 1979. Anomalous nutrient–chlorophyll interrelation-
ships in the offshore eastern tropical Pacific Ocean. J. Mar. Res.
37, 327 – 335.
Walsh, J.J., 1976. Herbivory as a factor in patterns of nutrient
utilization in the sea. Limnol. Oceanogr. 21, 1 – 13.
Welschmeyer, N.A., 1994. Fluorometric analysis of chlorophyll a in
the presence of chlorophyll b and phaeopigments. Limnol.
Oceanogr. 39, 1985 – 1992.
Wilkerson, F.P., Dugdale, R.C., Kudela, R.M., Chavez, F.P., 2000.
Biomass and productivity in Monterey bay, California:
contribution of the large phytoplankton. Deep-Sea Res. 47,
1003 – 1022.
Wright, S.W., Thomas, D.P., Marchant, H.J., Higgins, H.W.,
Mackey, M.D., Mackey, D.J., 1996. Analysis of phyto-
plankton of the Australian sector of the Southern Ocean:
comparisons of microscopy and size-frequency data with
interpretations of pigment HPLC data using the CHEM-
TAX matrix factorisation program. Mar. Ecol. Prog. Ser. 144,
285 – 298.