Iron has long been suspected to be a growth-limiting factor in the ocean based on well-established facts: its conversion to highly insoluble ferric hydroxide (rust) in alkaline, oxygenated seawater and its obligate requirement by all organisms. Furthermore, the consistently higher plankton productivity of near land compared with open ocean waters of the SO was taken as evidence for the provisioning of trace elements (including iron) from the land and their limiting role away from it (Hart 1942).
The perennially high macronutrient concentrations in large tracts of ocean, the Sub-Arctic and Equatorial Pacific and the entire SO, known as the HNLC regions, were considered a paradox compared with the equatorial and high-latitude regions of the Atlantic Ocean where nutrients were exhausted over much of the growth season. Three mutually inclusive reasons were proposed to explain the HNLC condition: low light levels in the deep mixed layers of high latitudes; iron limitation of phytoplankton growth; and heavier grazing pressure due to life histories of dominant grazers in HNLC regions. Although light limitation cannot apply to the Equatorial Pacific and zooplankton populations vary widely, both seasonally and regionally, iron limitation, the only factor common to all regions, was regarded with scepticism, partly because early iron addition experiments could show no difference between control and iron-supplemented bottles. With hindsight, this was due to contamination of controls by trace amounts of iron.
John Martin's group overcame the contamination problem by applying ultraclean methods to field measurements and bottle experiments. They demonstrated very low iron concentrations in HNLC waters and, in bottle experiments, a strong response of natural plankton to iron addition, in striking contrast to the absence of growth in control bottles (Martin & Fitzwater 1988). This evidence failed to convince many in the bio-oceanography community but biogeochemists took up the challenge more readily and have since opened a booming marine trace metal field (Jickells et al. 2005). Not surprisingly, most of the OIF experiments focused on biogeochemical processes with much less attention paid to the underlying ecological processes unfolding at the species level.
The reluctance of bio-oceanographers to place iron limitation of phytoplankton growth on an equal footing with light or macronutrients can be attributed to several reasons. Thus, the issue of whether N or P is the primary limiting nutrient in oceans relative to lakes (where P has long been accepted as the limiting nutrient) was hotly debated for both coastal eutrophic waters as well as the open ocean. However, the concept was not transferred to iron by the mainstream community, possibly because many of the scientists involved were working in coastal areas where iron was not an issue. Indeed an early, albeit inadvertent, iron fertilization experiment was carried out by a titanium factory in the form of acid waste dumping in the North Sea where no noticeable effects on phytoplankton productivity were reported. The practice was stopped due to popular protest and the acid waste is now converted to FeSO4 and applied to lawns and sewage treatment plants.
No doubt the methodological problems associated with contamination-free measurements of iron as well as maintenance of iron-clean conditions for experimentation have deterred many biological laboratories from studying the role of iron. Lack of clarity regarding the sources and availability of iron to offshore phytoplankton - how much of it is derived from upwelling from the deep sea vis-avis the amount introduced by dust (Cassar et al. 2007); and how much is retained in the surface layer by binding to organic molecules with an affinity for iron (ligands; Jickells et al. 2005) - is another factor hindering the acceptance of iron limitation by mainstream bio-oceanographers. Thus, the sources of iron enabling the annual spring bloom in the open North Atlantic, which set this ocean apart from the other high-latitude oceans, are still not unequivocally established.
Furthermore, there are indications that the iron demand of coastal versus oceanic phytoplankton differ (Strzepek & Harrison 2004) implying that the impact of iron deficiency, unlike that of macronutrients, is also a question of adaptation, hence species composition. However, this view overlooks the extensive, obligate role played by iron in the cellular machinery where it is as much an essential building block as, for example, phosphate. Both chlorophyll synthesis and nitrate reduction, the two gateways to light and macronutrient usage, respectively, require iron. Thus, by increasing cellular chlorophyll, iron-sufficient phytoplankton can shade adapt, thereby improving the efficiency of light usage (Gervais et al. 2002). The accompanying disadvantage of suffering photodamage under calm sunny conditions is of minor importance in the cloudy, stormy SO.
An example for the reluctance to accept the limiting role of iron is encapsulated in a sentence from a recent review of iron fertilization experiments (de Baar et al. 2005): 'Hence having dumped a total of 8975kg of Fe into HNLC waters and using approximately 1 year of shiptime, we may conclude that light is the ultimate determinate of the phytoplankton biomass response.' This statement is based on a comparison of changes in the concentrations of dissolved and particulate properties measured in the various experiments that vary with depth of the mixed layer. The lower chlorophyll concentrations induced by OIF in deeper water columns are interpreted as light limitation. However, budgetary analyses are based not on concentration (g C m-3) but on the magnitude of biomass, which is given by the total stock in the mixed layer water column: concentration multiplied by mixed layer depth (g C m-2). It is the magnitude of the stock (the amount of CO2 fixed in the water column) that determines the amount of CO2 exchanged between atmosphere and ocean and the amount that can sink to the deep sea. Indeed, the deeper the mixed layer, the larger the amount of biomass that can be built up, given sufficient iron, owing to the larger nutrient inventory of the deeper water column.
The dilution factor, which increases with depth of the mixed layer, has several different effects on the rate of biomass build-up and its eventual fate. Thus, the deeper the depth of mixing the smaller the percentage of the phytoplankton population within the euphotic zone (the layer where sufficient light is available to enable net growth), which will accordingly retard population growth rate (Smetacek & Passow 1990). The efficiency of light usage by phytoplankton will also be lower owing to attenuation by the deeper water column. However, these effects on the rate of biomass accumulation can be partly compensated by shade adaptation, which effectively increases the depth of the euphotic zone (Gervais et al. 2002), and by the reduction in grazing pressure due to the greater dilution of algal cells that decrease predator/prey encounter rate (Landry & Hassett 1982). Algal concentration could also affect the rate of aggregation of cells and chains into flocks (marine snow) and hence the magnitude of export from the surface and sinking rate through the deep water column. However, the relevant threshold concentrations of these effects need to be quantified in further experiments.
The fact that all SO OIF experiments induced blooms in a range of mixed layer depths and from spring to late summer (Boyd et al. 2007) indicates that iron availability and not light or grazing controlled the build-up of biomass. Thus, the standing stock of chlorophyll attained by the EisenEx bloom in a 70-m mixed layer after 21 days (200 mg Chl m-2) was the same as that reached by the SEEDS I bloom in the North Pacific following nitrate exhaustion after 13 days in a 10 m deep layer (Tsuda et al. 2003). Field observations also demonstrate that, apart from the winter months, light availability cannot be regarded as a limiting factor for phytoplankton biomass build-up. Standing stocks of more than 200 mg chlorophyll m-2, equivalent to that of North Sea spring blooms but in mixed layers three times as deep, have been recorded along the Polar Front (Bathmann et al. 1997) and continental slope (Turner & Owens 1995). Indeed, the standing stock of the latter diatom bloom of 7 mg Chl m-3 homogeneously distributed in a 70-m mixed layer, stretching in a band along the shelf break of the western Antarctic Peninsula, ranks among the highest recorded in the ocean. Such a standing stock (approx. 15gC m-2) could also be reached by OIF if the patch were large enough to prevent dilution with outside water. Clearly, this is a best-case scenario highlighting the need for more ambitious experiments.
Over the past decades, environmental-scale experiments carried out by terrestrial and lake ecologists have revealed unexpected fundamental insights on various relationships between ecological and biogeochemical processes. So, when the first in situ OIF experiments IronEx I and II demonstrated that a patch of surface ocean, marked with sulphur hexafluoride (SF6), could be successfully manipulated and followed for a significant period of time (Martin et al. 1994; Coale et al. 1996), the feat was hailed as the transition of ocean ecology from an observational to an experimental science (Frost 1996). Nevertheless, the response since then has been muted.
If only 10 per cent of the ship time available to the relevant chemical, biological and geological oceanography disciplines had been dedicated to carrying out OIF experiments, we would now be discussing the results of at least 30 experiments, of which one or more would be large-scale, long-term, multi-ship international projects akin to the North Atlantic Bloom Experiment coordinated by JGOFS in 1988 (Ducklow & Harris 1993). Such an experiment would by now have quantified the fate of iron-fertilized bloom biomass, provided new insights on the reaction of bacteria and zooplankton communities and their predators to enhanced productivity and monitored the reaction of the deep-sea benthos to an enhanced food supply. Furthermore, insights into the much debated relationship between species diversity and ecosystem productivity could also have been effectively addressed by OIF. In short, we would have answers to the open questions raised in recent reviews of OIF (de Baar et al. 2005; Boyd et al. 2007; Lampitt et al., Chapter 8) and acquired many more unexpected insights on pelagic ecosystem structure and functioning and their impact on the deep sea and sediments.
Cost, pollution and expertise cannot be the reasons for the lack of enthusiasm for OIF experiments. Because the ship operates in much the same place, in situ experiments actually burn less fuel than conventional transect oceanography racing from station to station. Ferrous sulphate required is sold at low cost in garden shops to improve lawns for which the recommended dosage is 20 gm-2. The dosage required to fertilize a bloom is 0.05 g m-2 for a 50 m deep mixed layer. Clearly, hazardous impurities in the commercially available ferrous sulphate will be diluted to insignificant levels. Dispersing iron is straightforward and using SF6 as a tracer is no longer required as other easily monitored parameters such as photosynthetic efficiency, increasing chlorophyll and declining pCO2 accurately track the patch. Running the experiment is not very different from interdisciplinary observational oceanography. Clearly, the prerequisites for oceanographic experimentation have long been in place.
Was this article helpful?