Limnol. Oceanogr.

, 36(8), 199 1, 1879-l 885

0 199 1, by the American Society of Limnology and Oceanography, Inc.

Effect of iron on productivity and size distribution of

Antarctic phytoplankton
E. Walter Helbling, Virginia VillafaAe, and Osmund Holm-Hansen
Polar Research Program, Scripps Institution of Oceanography, UCSD, La Jolla, California 92093-0202

Abstract
In shipboard experiments, addition of Fe to samples from Antarctic shelf waters or from deep
waters close to the shelf break did not have any detectable effect on phytoplankton populations.
Fe addition to pelagic waters, however, increased Chl a concentrations by a factor of 4-7 times
during l-2 weeks of incubation and also resulted in a shift from a nanoplankton-dominated
population to one dominated by microplankton. If these shipboard experimental results are ex-
trapolated to in situ results following enrichment of Antarctic pelagic waters with Fe, there may
be some mitigation of the greenhouse effect caused by elevated CO, concentrations in the atmo-
sphere. Not only would the rate of primary production increase, but also the percentage of primary
production that is exported to deep water might be increased because of Fe favoring the growth
of microplankton. -

The combination of high inorganic nu- cells that only low standing stocks can be
trient concentrations and low phytoplank- maintained.
ton biomass and low rate of primary pro- In this paper we address three aspects of
duction in pelagic Antarctic waters has been the question of the importance of Fe in Ant-
termed the major paradox of the Southern arctic waters. Does addition of Fe to Ant-
Ocean (El-Sayed 1987). Martin et al. (1990b) arctic waters increase the in situ growth rates
have recently suggested that in pelagic wa- and(or) total standing stock of phytoplank-
ters this is due to Fe limitation. According ton? Is any such effect of Fe relevant to both
to their data, waters over the Antarctic con- coastal and pelagic waters? Does the con-
tinental shelf are much enriched in Fe, and, centration of Fe affect the species compo-
hence, Fe should not be affecting phyto- sition or cell size distribution of the phy-
plankton dynamics in or near coastal wa- toplankton crop?
ters. Holm-Hansen and Mitchell (199 1) and
Mitchell et al. (199 l), however, have point- Materials and methods
ed out that coastal waters are extremely All work was done during two cruises on
variable in regard to phytoplankton stand- RV Surveyor (NOAA) in January-February
ing stock and often have phytoplankton bio- 1990 and January-March 199 1. Water sam-
mass similar to or less than the average found ples were obtained in deep waters of the
in pelagic waters (0.5 pg Chl a liter-l; El- Drake Passage and in or near shelf waters
Sayed 1988). Mitchell and coworkers of- around Seal Island (Fig. 1).
fered an alternative explanation for the In the 1990 work, water samples were
presence of high nutrient concentrations and obtained from an inflatable Zodiac at least
low phytoplankton biomass in Antarctic 400 m from the ship. Polycarbonate flasks
waters-that deep mixing results in such low (2 liter) were rinsed in 6 N reagent-grade
mean irradiance received by phytoplankton HCl and then soaked in 1 N HCl for some
weeks before use. Working in the bow of a
slowly moving Zodiac, these polycarbonate
Acknowledgments
flasks were first rinsed 6-7 times with sur-
This work was supported in part by NOAA Coop- face seawater and then filled by immersing
erative Agreement NA90AA-H-AF020, which was part them in front of the moving Zodiac. The
of the Antarctic Marine Living Resources (AMLR) flasks were placed in a plastic bag and re-
Program, and in part by National Science Foundation turned to our van on the ship. Working in-
grant DPP 88-17635.
We thank the officers and crew of NOAA Ship Sur- side a plastic hood within our van, either
veyor for support and help during cruises in 1990 and FeSO, (concentrations as shown in Figs. 2
1991. and 3) or FeSO,-EDTA (2 nmol EDTA to
1879
1880 Helbling et al.

59 Samples from the cultures were taken ev-

nC
ery day by opening the bottles within the
van and quickly pouring out sufficient vol-
ume (5 0- 150 ml) for analysis of Chl a, which
was done by fluorometric techniques with
lo-50 ml (Holm-Hansen and Riemann
1978) and for fractionating the phytoplank-
ton into nano- (<20 pm) and micro- (> 20
pm) size fractions. Samples were also pre-
served with buffered Formalin for floristic
analysis by standard inverted microscope
techniques, for which 1O-50 ml were settled
(Utermiihl 1958; Reid 1983). After removal
of a sample, the bottle was again sealed and
62 returned to the incubator. During the in-
59 57 55 53
cubation period, nothing was ever added to
LONGITUDE (“W)
the sample bottles or inserted into the bot-
Fig. 1. Location of sampling sites. Stations A and tles.
B were sampled in January-February 1990. Stations
C, D, E, and F were sampled in January-March 199 1. Results
A and F are -400 m from Seal Island. The dashed
line shows the location of the 1,000-m depth contour. Data in Fig. 2 show that adding Fe to
Insert-map showing the area of work (square). water obtained close to Seal Island (station
A) had no significant effect on either rate of
each nmol Fe) were added to the test sam- synthesis of Chl a or on total crop size
ples with a P- 100 Pipetman. The bottles achieved. A Chl a concentration of 70 pg
were then wrapped with two layers of neu- liter-’ can be expected to deplete all avail-
tral-density plastic to attenuate light inci- able NO3 (Sakshaug and Holm-Hansen
dent on the cultures to 25% of ambient and 1986). It should, be noted that there was no
the flasks were then exposed to solar radi- discernible lag phase in any of these cul-
ation (resulting irradiance incident on the tures. The growth rate, assuming a constant
samples -300 PEinst rnp2 s-l) in a water- C : Chl a ratio, was -0.89 doubling per day,
cooled incubator on the deck of the ship for which corresponds to a specific growth rate
l-2 weeks. (h) of 0.62 d-l. There were no significant
In the 199 1 studies, water samples were differences in species composition in the
obtained from the ship’s continuous clean- control and experimental bottles; they were
water intake system which served scientific dominated by microsize diatoms, especially
purposes only. This system, which was by Chaetoceros spp. and Nitzschia spp.
“open” continuously while at sea, consisted The samples obtained from deep water
mainly of plastic tubing, but also contained just north of the continental shelf dropoff
a short piece of bronze tubing and some (station B) also showed no significant re-
stainless steel portions close to the intake. sponse to addition of Fe (Fig. 2). As the four
Microscopic observations as well as total bottles were individually filled, the initial
and fractionated Chl a analyses of samples Chl a concentrations varied from 0.47 to
taken from this system showed no differ- 0.68 pg liter- I. The growth rates of the four
ences compared to 5-m water samples ob- cultures were all -0.57 doubling per day (p
tained with Niskin bottles. After additions = 0.39 d-l), which was slower than that of
of FeSO or FeSO,-EDTA to the 2-liter, the cultures shown for station A (Fig. 2).
acid-cleaned polycarbonate bottles, all sam- There were no significant differences in phy-
ples were placed in a refrigerator equipped toplankton species composition in these four
with daylight fluorescent lamps. Tempera- cultures. In contrast to the experiment from
ture was maintained between 0” and 4°C and coastal waters, the phytoplankton crop in
the irradiance regime was 12 : 12 L/D (150 these offshore waters was dominated by
PEinst rnd2 s-l). small Nitzschia spp. of the Fragilariopsis
 Efect of Fe on phytoplankton 1881

ni
3 4

0 12 3 4 5 6 7 8 9 10
Days
Fin. 2. Effect of Fe addition on the rate of synthesis of Chl a. Concentrations of Fe shown are expressed in
nmoi Fe liter-l. Data collected in January 1996.

section (e.g. N. pseudonana and N. cylin- trol bottles. The response of Fe addition to
drus), small flagellates < 15 pm in diameter, waters from station D was similar to that
and larger Nitzschia spp. (30-39 pm). obtained at station C, both in regard to Chl
The 199 1 sample from mid-Drake Pas- a concentrations (Fig. 3) and to relative
sage waters (station C) showed pronounced amounts of nano- and microplankton. Mi-
differences between the controls and the + Fe croscopic observation of the samples from
treatments (Fig. 3). All treatments were quite station C during the exponential phase of
similar with positive but slow growth for growth showed that in the +Fe treatment
the first 2-3 d, after which the +Fe bottles there was a large increase in concentration
showed exponential growth to -20 pg Chl of cells when compared with the controls.
a liter-l, whereas the controls leveled off at In the controls, small diatoms (Nitzschia
- 3. The size distribution of the phytoplank- spp.) and small flagellates (< 10 pm) were
ton crops in the different treatments also dominant. These small cells also increased
changed significantly. The + Fe treatments in the +Fe treatments, but microsize cells
favored the growth of microsized ‘species, (e.g. Rhizosolenia hebetata fo. semispina,
so that by the end of the growth period, the Corethron criophilum, and pennate diatoms
nanoplankton fraction in the +Fe bottles -40-50 pm long) accounted for most of the
for station C averaged -20% of the total increase of biomass.
biomass as judged by Chl a concentrations, Fe additions to water samples obtained
compared to the average of 65% for the con- in 1991 from shelf waters or north of the
1882 Helbling et al.

\ .! .)...
1: Station C “i : i. : :” 100
Station D ?:I
. . ,..

IO

-o-CONTROL

0.1 1
0 2 4 6 8 IO 12 14 16 0 2 4 6 8 10 I2
_ ., . ,
:. . Station E I I IO0 : Station F f.
I
.,
>.

: Jzp- ,‘P’ :
I-
m-‘. ,: , I

: : : 1
t Q CONTROL1

I e-CONTROL
r-Fe 5
-.Fe 5
ni
0 2 4 6 8 IO 12 14""o 2 4 6 8 10 12 14 16
7

$20
10
n -1

"0 2 4 6 8 IO 12 14 -0 I 2 3 4 5 6 7 8 9
Days Days
Fig. 3. Effect of Fe addition on the rate of synthesis of Chl a and on size distribution of phytoplankton based
on Chl Qmeasurements. Concentrations of Fe shown are expressed in nmol Fe liter-l. Data collected in January-
February 1991.

shelf break (stations F and E) did not show plankton crop (Fig. 3). These results are
any influence on either growth rates, Chl a similar to those obtained during the 1990
concentrations at the end of the growth pe- studies, even to the extent that both stations
riod, or on size distribution of the phyto- E and B supported -20-30 pg Chl a liter-‘,
 Eflect of Fe on phytoplankton 1883

while stations F and A supported -70 yg in deep water, they are close to the shelf
Chl a liter-l. dropoff, and it can be expected that currents
from the Bransfield Strait and Bellingshau-
Discussion sen Sea will transport high Fe-containing
In view of the precautions exercised by waters north to these locations.
other researchers to minimize metal con- Addition of Fe to waters farther north of
tamination in studies of Fe in marine wa- the continental shelf (stations C and D) does,
ters, it is perhaps surprising that our results however, have pronounced effects on both
duplicated results from investigations with the concentrations of Chl a achieved at sta-
“clean” sampling techniques (Martin et al. tionary phase as well as on the size distri-
1990a; de Baar et al. 1990), in spite of the bution of the phytoplankton crop. It should
fact that some of our water samples were be noted that the effect of Fe addition is not
obtained from the ship’s clean intake sys- immediate, but takes 2-4 d to become ev-
tem. Our data, however, suggest that metal ident in regard to Chl a concentrations (Fig.
contamination (either essential or toxic el- 3, stations C and D). The results of Mar-
ements) did not occur in our samples to any tin and Fitzwater (1988), de Baar et al.
significant extent. (1990), and Martin et al. (1990a) often show
Our evidence is as follows. Samples from a similar lag of 2-4 d in Chl a synthesis in
shelf waters, or just north of the shelf break, both control and +Fe treatments in waters
gave similar results when samples were ob- low in dissolved Fe. Such a lag in Chl a
tained either from the ship intake system or synthesis after addition of a limiting nutri-
from a Zodiac far from the ship. In shelf ent has been documented in many labora-
waters, there was no effect on growth rates tory cultures of marine phytoplankton (e.g.
or on biomass achieved by adding Fe or Holm-Hansen 1970). Thus the lag noted in
EDTA, regardless of how the water sample many Fe experiments in natural waters is
was obtained. There was no significant lag suggestive of nutrient limitation by Fe, but
in cell growth (based in Chl a concentra- it could also be caused by grazing effects or
tions) in any of our experimental samples. death of some cells after containment in
In offshore waters, samples with Fe addi- bottles (see Banse 199 1, for discussion on
tions showed more biomass than the control this point).
samples only after 3-4 d, which is the length One of the most difficult questions to an-
of time generally noted by Martin et al. swer regarding the role of Fe in pelagic Ant-
(1990a) and de Baar et al. (1990). The bio- arctic waters is whether Fe is actually af-
mass values achieved in our +Fe and con- fecting in situ phytoplankton growth rates.
trol samples from offshore waters were com- Many attempts to document nutrient lim-
parable to those reported in the Fe studies itation in Antarctic water have failed (e.g.
mentioned above. One explanation for the Hayes et al. 1984), resulting in a general
lack of noticeable contamination in our view that phytoplankton populations in
samples from the ship intake system was Antarctic waters are not limited by nutrient
that this system was open at all times while deficiencies. The apparent contradiction of
at sea; hence the continuous flow minimized these data to the Fe-addition experiments
the concentrations of any metals leached can be explained, however, by the fact that
from the tubing. the older nutrient limitation experiments
Our data suggest that Fe limitation does generally measured radiocarbon incorpo-
not have any detectable effect on Antarctic ration in incubations of I 1 d. If the samples
phytoplankton in or close to shelf waters. are from coastal waters, all available data
The lack of any lag phase (Fig. 2) suggests suggest that nutrient limitation is of no sig-
that these phytoplankton populations were nificance; if the water samples are from pe-
not nutrient limited at the time of sampling. lagic waters, then examination of existing
This is in agreement with the data of Martin data (Martin and Fitzwater 1988; Fig. 3,
et al. (1990b) showing that shelf waters have stations C and D) shows that there is gen-
high concentrations of Fe compared to pe- erally no discernible difference between
lagic waters. Although stations B and E are controls and Fe treatments in regard to ei-
1884 Helbling et al.

ther NO3 uptake or Chl a concentrations for size distribution of phytoplankton. The ef-
l-3 d. Presumably the rate of CO2 uptake fects on cell size could have a feedback on
during the first few days of the experiment grazing rates by zooplankton because of size
also would show no significant differences preference of food particles, which might in
between treatments. turn have an effect on the amount of pri-
The above comments apply to experi- mary production that is exported to deep
ments dealing with the “total” phytoplank- waters via fecal pellets.
ton crop in natural water samples. Adding Such possible effects of Fe addition on the
Fe to pelagic waters, however, favors the C cycle assume that it is valid to extrapolate
growth of microsize phytoplankton (Fig. 3, our experimental results from deck incu-
station C), as might be expected from the bation to events that would occur if Fe were
Fe-assimilation studies of Hudson and Mo- added directly to pelagic Antarctic waters.
rel (1990). Such growth has much potential To deplete most of the available NOls in
significance as the phytoplankton crop in Antarctic surface waters would require that
most pelagic waters of the Antarctic con- high phytoplankton growth rates be main-
tains 75-95% nanoplankton (Hewes et al. tained for l-2 weeks, as occurred in our
1990; von Briickel 198 I). If the addition of incubation experiments, which would be so
Fe increases the growth rates of the larger only if no other factor came into play to
microplankton as compared to growth rates limit phytoplankton growth rates. The two
of nanoplankton, then a consequence of the major factors to consider in this context are
very low standing stock of microplankton possible limitation of growth by some other
in pelagic waters would result in a consid- essential micronutrient (e.g. Mn, MO, Zn,
erable lag phase before the consequences of or Co) and available solar radiation.
Fe addition become evident from total Chl Mitchell et al. (199 1) have modeled the
a measurements. This scenario would im- relationship between rates of primary pro-
ply that the nanoplankton population either duction and the depth and stability of the
is not as nutrient stressed as the microplank- upper mixed layer and concluded that phys-
ton or that its biomass is controlled by the ical mixing processes in the pelagic portions
ubiquitous heterotrophic protozoan popu- of the Southern Ocean are the major factors
lations as described by von Brijckel (198 1) involved in preventing phytoplankton
and Hewes et al. (198 5). It is seen from Fig. blooms. They pointed out, however, that
3 that the controls at station D supported the estimates of primary production pre-
higher crops (6-10 pg Chl a liter-‘) than at dicted by this model are highly dependent
station C (- 3 pg Chl a liter-‘). It is likely on loss factors such as grazing, cell settling,
that this is due to higher Fe concentrations and respiration. Because these loss factors
near the continental shelf and progressively are largely dependent on species composi-
lower concentrations in pelagic waters as tion and cell size, our experimental results
one gets farther from the Fe-rich shelf wa- indicating that Fe additions may enhance
ters. the biomass of microplankton compared to
The addition of Fe to Antarctic waters that of nanoplankton should be taken into
has been suggested (Martin et al. 1990a,b; account when considering Mitchell et al.‘s
but see Davies 1990) as a means to sequester predictions of the light control model of
large amounts of CO* from the atmosphere phytoplankton growth.
and thus to mitigate the consequences of the tt is thus of importance to examine the
greenhouse effect resulting from increasing significance of Fe-induced changes in flo-
concentrations of CO2 in the atmosphere. ristic composition on the photobiological
Although our data do not provide an answer characteristics (e.g. PI parameters) of the
to the question of whether Fe is affecting in phytoplankton population and also to de-
situ phytoplankton growth rates, they do termine whether Fe additions are affecting
suggest that Fe additions to pelagic waters the physiological properties of any portion
may significantly increase the magnitude of of the phytoplankton crop during the 2-3 d
the phytoplankton crop and also alter the immediately following addition of Fe. It
 Ecfect of Fe on phytoplankton 1885

would seem expedient to resolve these ques- adenosine triphosphate and particulate organic
tions before implementing any plan for large: carbon. Mar. Ecol. Prog. Ser. 63: 27-35.
HOLM-HANSEN, 0. 1970. ATP levels in algal cells as
scale fertilization of the Antarctic Ocean. influenced by environmental conditions. Plant Cell
Physiol. 11: 689-700.
References AND B. G. MITCHELL. 199 1. Spatial alrd tem-
BANSE, K. 199 1. Iron availability, nitrate uptake, and poral distribution of phytoplankton and primary
exportable new production in the subarctic Pacific. production in the western Bransfield Strait region.
J. Geophys. Res. 96: 741-748. Deep-Sea Res. 38: 96 l-980.
BR~CKEL, K. VON. 198 1. The importance of nano- -, AND B. RIEMANN. 1978. Chlorophyll a de-
plankton within the pelagic Antarctic ecosystem. termination: Improvements in methodology. Oi-
Kiel. Meeresforsch. Sonderh. 5: 6 l-67. kos 30: 438-447.
DAVIES, A. G. 1990. Taking a cool look at iron. Na- HUDSON, R. J. M., AND F. M. M. MOREL. 1990. Iron
ture 345: 114-l 15. transport in marine phytoplankton: Kinetics of
DE BAAR, H. J. W., AND OTHERS. 1990. On iron lim- cellular and medium coordination-reactions. Lim-
itation of the Southern Ocean: Experimental ob- nol. Oceanogr. 35: 1002-l 020.
servations in the Weddell and Scotia Seas. Mar. MARTIN, J. H., AND S. E. FITZWATER. 1988. Iron de-
Ecol. Prog. Ser. 65: 105-122. ficiency limits phytoplankton growth in the north-
EL-SAYED, S. Z. 1987. Biological productivity of the east Pacific subarctic. Nature 331: 341-343.
Antarctic waters: Present paradoxes and emerging - - AND R. M. GORDON. 1990a. Iron
paradigms, p. l-2 1. In S. Z. El-Sayed and A. Tomo dekciency limits phytoplankton growth in Ant-
[eds.], Antarctic aquatic biology. SCAR, Cam- arctic waters. Global Biogeochem. Cycles 4: 5-l 2.
bridge. -, R. M. GORDON, AND S. E. FITZWATER. 1990b.
-. 1988. Seasonal and interannual variabilities Iron in Antarctic waters. Nature 345: 156-l 58.
in Antarctic phytoplankton with reference to krill MITCHELL, B. G., E. A. BRODY, 0. HOLM-HANSEN, C.
distribution, p. 101-l 19. In D. Sahrhage [ed.], MCCLAIN, AND J. BISHOP. 199 1. Light limitation
Antarctic Ocean and resources variability. Spring- of phytoplankton biomass and macronutrient uti-
er. lization in the Southern Ocean. Limnol. Oceanogr.
HAYES, P. K., T. M. WHITAKER, AND G. E. FOGG.
36: 1662-l 677.
1984. The distribution and nutrient status of phy-
toplankton in the Southern Ocean between 20” and REID, F. M. H. 1983. Biomass estimation of com-
7O”W. Polar Biol. 3: 153-165. ponents of the marine nanoplankton and pico-
HEWES, C. D., 0. HOLM-HANSEN, AND E. SAKSHAUG. plankton by the Utermijhl settling technique. J.
1985. Alternate carbon pathways at lower trophic Plankton Res. 5: 235-252.
levels in the Antarctic food web, p. 277-283. In SAKSHAUG, E., AND 0. HOLM-HANSEN. 1986. Pho-
W. R. Siegfried et al. [eds.], Antarctic nutrient toadaptation in Antarctic phytoplankton: Varia-
cycles and food webs. Springer. tions in growth rate, chemical composition and P
-, E. SAKSHAUG, F. M. H. REID, AND 0. versus I curves. J. Plankton Res. 8: 459-473.
HOLM-HANSEN. 1990. Microbial autotrophic and UTERM~HL, H. 1958. Zur Vervollkommnung der
heterotrophic eucaryotes in Antarctic waters: Re- quantitativen Phytoplankton-Methodik. Mitt. Int.
lationships between biomass and chlorophyll, Ver. Theor. Angew. Limnol. 9. 38 p.