Annex B. Sequestration of carbon dioxide by ocean fertilisation.

1. INTRODUCTION

Currently, the worlds oceans store about 40,000 Gt and act as both a large source and sink of atmospheric CO₂. Almost 90 Gt C is naturally cycled between the ocean and the atmosphere every year. It is assumed that prior to industrial revolution there was a zero net flux of CO₂ between the two reservoirs, and that the net flux to the ocean from the atmosphere is therefore driven by anthropogenic CO₂. Whether different regions of the oceans act as a source or sink of CO₂ depends on whether surface seawater concentrations are greater or lower than the atmosphere after allowing for solubility considerations.

The main route by which CO₂ enters the deep oceans is at high latitudes where surface water cools before sinking. As the solubility of CO₂ increases with decreasing temperature, these cooler waters absorb atmospheric CO₂. This cycling mechanism is known as the solubility pump. In contrast, deep waters upwelling to the surface in tropical regions contain high levels of CO₂ and therefore act as sources to the atmosphere. Increasing levels of atmospheric CO₂ over recent times mean that the ocean uptake in high latitude regions will have increased.

Transfer of CO₂ from the surface oceans to the deep water is also brought about by photosynthesis (biologically pumped - see figure 1). Marine algae are responsible for nearly one third of the global gross photosynthetic production. The important fraction of this cycling is the amount of carbon that is lost from the surface layer to the deeper ocean (the export production) compared to the carbon that is simply recycled in, or near, the euphotic zone (up to the top 50 metres, see figure 1). It seems unlikely that there has been any increase in the biological pump with increasing atmospheric CO₂ as only under rare conditions is carbon likely to limit algal growth in the oceans.

There is little doubt that the deep oceans have the capacity to absorb almost all of the anthropogenic CO₂ presently in the atmosphere. This is due to the so called buffering of seawater or its alkalinity. Inorganic carbon is partitioned between three main chemical species in the oceans, dissolved CO₂, bicarbonate ions (HCO₃^-) and carbonate ions (CO₃^2-). For every 100 units of inorganic carbon, approximately 98 are present as HCO₃^-, one as CO₃^2- and one as CO₂ e.g. an addition of an extra 100 units CO₂ to seawater would result in only extra one unit remaining as CO₂.

The reason why the oceans have not absorbed most of the atmospheric anthropogenic CO₂ lies in the time-scale involved in equilibrating the deep oceans with the atmosphere. The rate-limiting step is the transfer of water between the surface and deep oceans, the equilibration time of which is several thousand years. Most realistic geo-engineering schemes for ocean sequestration of man-made CO₂ are based on increasing the rate of transfer of carbon from the atmosphere (or from industrial point sources) to the deep oceans. Sequestration of anthropogenic CO₂ by the deliberate fertilisation of the ocean can only work if either the solubility or the biological pumps, or both, are increased. The ocean fertilisation idea works on the premise that the rate of drawdown of CO₂ in the biological and the solubility pumps can be increased.

Figure 1. Schematic of the ocean carbon biological pump (from Houghton et al. 1996)

2. FERTILISING THE OCEAN WITH IRON

2.1 Description

The presence of "high-nutrient low-chlorophyll" (HNLC) surficial waters in one-fifth of the worlds oceans has been attributed to iron limitation, although other explanations such as insufficient light or micro-nutrient availability and grazing pressure² have also been considered. Iron is only soluble in its reduced form and this restricts it availability in the ocean; its presence in the surface waters is largely attributable to atmospheric dust transport. The "iron hypothesis" assumes that phytoplankton utilisation of the macro-nutrients nitrate, phosphate and silicate is effectively "capped" in the remote HNLC regions due to the limited supply of iron. Cessation of iron limitation in bottle experiments and three international mesoscale iron release experiments (IRONEX, IRONEX2, and SOIREE) in the three main HNLC regions (the Equatorial Pacific, the Sub-Arctic Pacific and the Southern Ocean) have confirmed the premise that iron is limiting phytoplankton production in these regions. These observations have stimulated debate as to whether the addition of iron to the surface waters of HNLC regions could enhance carbon export to the deep ocean and decrease atmospheric CO₂.

2.2 Previous and on-going research

Support for the theory of iron limitation over long time scales was obtained from the Vostok ice core, in which inferred elevated iron levels correlated with low atmospheric CO₂ during glacial periods⁸. The coincidence of high atmospheric dust levels derived from the Vostok core and elevated organic carbon accumulation in deep ocean cores provided further circumstantial evidence⁹. In addition, there is indirect evidence that Southern Ocean productivity increased temporarily following the Mt. Pinatubo eruption, perhaps as a result of iron fertilisation by dust transported via the stratosphere¹⁰. Further circumstantial evidence was obtained from the correspondence of iron-mediated phytoplankton blooms and export pulses of particulate organic matter to the deep water¹⁷.

Direct evidence of the iron limitation of phytoplankton has been provided by iron addition to shipboard incubations of Antarctic and NE subarctic waters³, although the interpretation of these results has generated considerable debate. The exclusion of factors such as mixing and grazing by herbivorous zooplankton may generate artefactual results in these incubations, and may account for the phytoplankton increases observed in the untreated "control" incubations. Useful as such experiments are, conditions in the open sea can never be reproduced in vitro, and there will always be considerable uncertainty when extrapolating from small-scale incubations to the "real" system.

The two IRONEX experiments in the Pacific Ocean proved unequivocally that addition of iron increased productivity⁴ in this region. The addition of 450 kg of dissolved iron to an area of approximately 60 km² in the Equatorial Pacific increased iron concentrations by a factor of 40 and stimulated a fourfold increase in phytoplanktonic production rates and an associated decrease in surface CO₂. However the effect was short-term and rapid recycling of the extra production by the grazing and microbial components of the ecosystem occurred after 4 days⁴. However, in the second experiment, an addition of ~4 nmol l^-1 of dissolved iron, equivalent to a 10 to 20 fold increase in iron concentration, triggered a dense diatom bloom which resulted in a 30-fold increase in productivity and biomass over a 3 week period. Significant increases in nutrient uptake were recorded with a corresponding decrease of 60% in surface CO₂⁴. The success of the second experiment was attributed to the addition of iron in successive infusions, as opposed to one single release, thereby reproducing natural supply of iron to surface waters. However, limitation by light may have depressed the response in IRONEX1 following subduction of the fertilised patch to a depth of 30m. However, as a result of these experiments we are now certain that iron limitation plays a fundamental role in the biology and chemistry of the equatorial Pacific.

The extent to which any stimulation of biological activity will increase carbon sequestration into the deep ocean is dependent upon the local circulation and hydrography. The Equatorial Pacific, where the initial experiments were performed, is characterised by strong upwelling of subsurface waters with lateral transport to sub-tropical regions, and carbon export to the deep is effectively minimised. In contrast, transport between surface waters and the large mass of subsurface water in the Southern Ocean where the majority of the "labile" biospheric carbon resides is relatively rapid⁵. The polar oceans therefore function as gateways between the atmosphere and the deep sea, with the Southern Ocean being the only polar ocean that maintains a large pool of unused nutrients.

As a result carbon cycle models suggest that it is predominately the Southern Ocean which maintains the natural atmospheric levels of CO₂ and that changes in iron availability in the geological past may have influenced glacial-interglacial variation in CO₂ ⁶. Furthermore it is clear that it is the rate of vertical mixing in the Southern Ocean which is the key factor in determining the response to iron fertilisation.

Application of a box model of fertilisation of Antarctic surface waters concluded that after 100 years of successful fertilisation, the atmospheric CO₂ content would be lowered by 50 +/- 25 ppm which would represent approximate 10 +/-5 % of the atmospheric total¹⁶. A further study used a 3-D global ocean model to calculate an upper limit to the oceanic uptake of CO₂ that would result from complete biological utilisation of surface water macro-nutrients in the Southern Ocean⁷. In the business-as-usual scenario, the oceanic carbon uptake without macro-nutrient depletion is predicted to increase over the 100 year period from 2 Gt C yr^-1 to 6 Gt C yr^-1. With total depletion of Southern ocean nutrients this uptake is predicted to be enhanced by 1-2 Gt C yr^-1 with the atmospheric CO₂ predicted to be 72 ppm lower at the end of the 100 year period than without any macro-nutrient depletion. This amounts to the removal of an additional 152 Gt C from the atmosphere. Although this is a large effect, it is still less than the increase in atmospheric CO₂ resulting from business-as-usual emissions (an increase of 430 ppm). Under a constant emission scenario, a similar increase in uptake is predicted (which therefore has a proportionally larger effect). A general summary of the present literature on modelling of iron fertilisation of the Southern Ocean concludes that it could potentially lower atmospheric CO₂ by 6-21%^4,7,20.

The 3d modelling study demonstrates that there would be a large oceanic uptake of CO₂ as a result of complete macro-nutrient depletion in the Southern Ocean. On its own, this would not provide a solution to the problem of global warming but could make a significant contribution to reducing it. However, complete nutrient depletion of the Southern Ocean is an extreme scenario which would be difficult to achieve and would cause potentially serious ecological risks.

The phytoplankton taxon response to the iron may have a critical effect on the relative effectiveness of any iron fertilisation. Stimulation of the diatoms, which are predominately organic with minimal calcite carbon content, would alter the inorganic:organic carbon content of the detrital material exported to the sediment. This may lead to a decrease in the pH of the surface sediment and increased dissolution of calcite. On a longer timescale this could result in an increase in alkalinity and lowering of CO₂, so that the air-sea CO₂ gradient would be greater upon upwelling, so further enhancing carbon uptake¹⁰.

Recently the first in situ iron fertilisation experiment (SOIREE) took place in the Australasian section of the Southern Ocean, during which dissolved iron levels were maintained at or above 10 times background levels for a period of 13 days. The response to the iron was slower than in the Pacific experiments as a result of lower physiological growth rate in the colder waters and also the more limited light availability in the deeper surface mixed layer of the Southern Ocean. Enhancement of phytoplanktonic growth rates, biomass and photosynthetic efficiency and decreased surface CO₂ were apparent within the fertilised patch on comparison with external waters. At the time of writing the numbers have not been finalised but the results confirm that iron availability plays a fundamental role in controlling phytoplankton growth in this region of the Southern Ocean¹⁸.

2.3 Potential scale for approach by Governments

Present understanding of the carbon cycle suggests that the proportion of unused nutrients and associated carbon at the surface of the Southern Ocean is critical in setting the natural global atmospheric CO₂ concentration. Model simulations suggest that it is only in the Southern Ocean that enhanced phytoplanktonic production will cause a substantial change in atmospheric CO₂ concentrations⁷. Iron fertilisation does therefore have some potential to increase the oceanic uptake of CO₂ but, realistically, it can only contribute on a significantly smaller scale than that investigated in the above modelling work. Modelling does suggest that much of the effect could be obtained by concentrating the iron fertilisation effort in those areas of the Southern Ocean where the supply of nutrients to surface waters is greatest³. 78% of the increase in new production predicted to occur in 18% of the total Southern Ocean area, where convective plumes are present.

2.4 Likely environmental impact

In our present state of knowledge we cannot determine whether deliberate iron fertilisation in the Southern Ocean would be wasteful, or even counter-productive, due to uncertainties in our understanding of the overall response of the Southern Ocean ecosystem. There are potentially negative and deleterious ecosystem responses to iron addition which may exacerbate current global environmental problems.

An increase in vertical transport of organic matter production and subsequent microbial mineralisation may lead to a decline in the oxygen level of the water column, with generally negative implications for the biology and biogeochemistry. Modelling of the extent of deoxygenation predicts an overall depletion of only 10 micromol/kg compared with a mean concentration of 168 micromol/kg in the deep ocean. However, this is undoubtedly an underestimate as the mineralisation of the organic matter and subsequent oxygen decline will be more localised over depth and area.

From a biological viewpoint, deoxygenation of the water column would potentially lead to high mortalities across the biota and the development of "dead zones". At low oxygen levels there would be a significant reduction in overall biodiversity, and it is possible that even moderate oxygen decline would stimulate shifts in the species composition and an overall decline in biomass. The sessile and sedentary species within the sedimentary community would be most at risk from the effects of reductions in oxygen.

Increased mineralisation of the sinking organic matter may, under extreme circumstances, result in the complete removal of all available oxygen. This would result in a shift towards microbial reductive pathways altering the cycling and chemical speciation. In this case, the onset of denitrification in the absence of oxygen would lead to the loss of nitrate from the system with reduced nitrogen availability for phytoplanktonic uptake. Other effects of such a scenario would include increases in sulphate reduction with production of hydrogen sulphide, methanogenesis (see below) and increased mobilisation of iron with an increase in the dissolved reduced Fe (II) phase.

Iron addition will impact upon the source strengths of other radiative gases, such as methane (CH₄₎ and nitrous oxide (N₂O) via a shift towards hypoxia. Although oxygen-deplete environments are classically regarded as the significant sources of these gases, oxygen rich waters are generally supersaturated with respect to these gases¹². It would seem logical that bacterial process rates and the microsites which support these processes will increase in the surface mixed layer in response to the increased particulate carbon and heterotrophy resulting from iron fertilisation. The enhanced particle export will lead to increases in rates of methanogenesis, denitrification and nitrification in the mid and deep water, and consequently will increase production of both methane and nitrous oxide gases. However, the transport of these biogases to the atmosphere would occur over relatively long time scales.

Changes in the source strength and resultant air-sea fluxes of these gases has not been examined during previous iron fertilisation experiments despite their potential to negate any benefit derived by increasing oceanic CO₂ uptake. On a molecular basis, both gases are significantly stronger greenhouse gases than CO₂, with Global Warming Potentials (GWP) of 62 and 290 times that of CO₂ estimated on a 20 year time horizon¹⁴. The enhancement of N₂O production in particular represents a major potential drawback to the use of iron fertilisation to mitigate the effects of global warming. It has been estimated that the decrease in GWP weighted emissions resulting from carbon removal as CO₂ upon iron fertilisation would be completely counteracted by the N₂O produced by microbial denitrification and nitrification¹⁴. This may be an overestimate as it is based upon maximum rates, but does not include N2O production within the euphotic zone. A second more realistic approach, directly relating N2O production to new phytoplankton production, suggests that the additional N2O would account for 50-95% of the greenhouse potential of the carbon removed by ion fertilisation¹⁴. As enhanced N₂O production might largely offset any mitigation of global warming achieved by increasing carbon sequestration in deliberate iron fertilisation scenarios, a "biogas budget" is clearly desirable to determine the integrated impact of iron fertilisation upon climate change.

In IRONEX2 there was a general increase in biomass across the phytoplanktonic community with most taxa responding to addition of iron. In the SOIREE experiment the smaller phytoplankton initially responded to the iron with larger species dominating in the latter stages. Change in phytoplankton speciation and community dominance will not only impact further up the food chain, but will also influence the biogeochemistry through various chemotaxonomic routes. The gas dimethylsulphide (DMS) is derived from dimethylsulphoniopionate (DMSP) which is to balance cell water content by certain phytoplanktonic taxa. Both DMS and DMSP exhibited three-four fold increases in the iron-fertilised waters during IRONEX2², with a larger increase observed in the SOIREE experiment in the Southern Ocean (S. Turner, pers.comm.). DMS contributes to the non-seasalt sulphur in the atmosphere, and oxidation is linked to cloud formation by the provision of condensation nuclei (CCN)¹¹. However, this is largely inferred and the connection between DMS and CCN has yet to be established conclusively. Furthermore, whereas increased cloud cover may increase reflectance, and so potentially lower atmospheric temperature, it may also increase the water vapour content of the atmosphere and so the overall greenhouse forcing.

Halocarbon measurements during IRONEX1 identified an increase in dissolved methyl iodide⁴, a trace atmospheric gas which influences atmospheric hydroxyl content and contributes slightly to ozone depletion in the stratosphere. This compound is considered to be produced intra-cellularly by phytoplankton, and so has potential implications for atmospheric chemistry in the event of stimulated phytoplankton production in response to iron addition. Finally, it should also be considered that the development of toxic blooms of algae may be stimulated by iron fertilisation, as the appearance of these in other oceanic regions has been attributed to enhanced nutrient status and eutrophication.

It should also be noted that a serious hurdle would be public perception of iron fertilisation, particularly with respect to the Antarctic waters. As this is regarded as the last pristine environment and is currently protected by a number of international treaties and moratoria, iron fertilisation of this region may still be unacceptable even if the benefits can be clearly demonstrated and ecosystem and aesthetic impacts shown to be minimal.

2.5 Effect on Ocean Productivity

An increase in the new production 12-14 Gt C yr^-120 was estimated for the Southern Ocean which represents a doubling of the estimated present day global new production. This increase in new production would be expected to continue indefinitely as there is a continual upward supply of regenerated nutrients. This is a significant change that would undoubtedly be reflected by increases in the zooplankton population and stocks of fish and marine mammals. However this increase in the biota would not occur without significant changes in species composition at each level of the food chain. Outside the Southern Ocean the model predicts that new production may be reduced by 2-3 Gt C yr^-1. This has been explained by the nutrient supply in equatorial regions being from upwelling rather than convective mixing, as in the Southern Ocean.

It is clear that the microbial fixation of nitrogen is a significant source of nitrogen for the ocean and that this process is predominately limited by iron availability¹⁹. As it has been shown that fixed nitrogen limits phytoplankton production on geological timescales, this suggests that stimulation of nitrogen fixation upon addition of iron would further enhance productivity and export of the Southern Ocean, and potentially oligotrophic oceanic regions in the tropics.

2.6 Economic costs

The cost of any iron fertilisation would largely derive from distribution of the iron and maintenance of dissolved levels over significant time and space scales. Upwelling of major nutrients in the Southern Ocean could potentially support 1.8 billion tons C per annum. Estimates of the amount of Fe required to fertilise this region (during austral spring and summer) range from 1-5 x 10⁵ tons depending upon the Fe:C ratio¹⁵. The economic cost would be further offset by the release of carbon dioxide (and other gases) associated with the extraction and pre-treatment processes and transport of the iron to the Southern Ocean and its subsequent distribution.

Table 1 shows estimates of the practical requirements for Fe fertilisation²⁰ based upon the earlier models of Sarmiento and Orr, 1991³. These figures assume that 37% of the decrease in atmospheric CO₂, or 78% of the new production, could be achieved by fertilising the convective plume regions which account for just 18% of the Southern Ocean.

Table 1. Logistics for Iron Fertilisation
New production	10.3 Gt C per annum
Carbon sequestered	0.5 Gt C pr annum
Fe required (assuming C:Fe of 105:1)	0.47 Mt per annum
Area of ocean	1.7 x 107 km2
Travel distance for application	8.8 x 108 km per annum
No of ships required (at 20 knots)	2700 (pa)
No. of aircraft (at 600 km/hr)	200 (pa)

Separate estimates of the cost of iron fertilisation put the combined cost of obtaining and distributing the iron by ship at between £3 and £37/tC²¹.

2.7 Practicality

The technology to achieve fertilisation of the Southern Ocean would be on a relatively low level of complexity and cost to design and implement. Dynamic release of the iron from mobile distribution sources such as ships or planes would not require significant technological development. Table 1 suggests that airborne application of the iron may provide the most economical approach to distributing the iron over such a large area. The slow release of iron from floating pellets suggests potential but requires further research and investment.

2.8 Likely timescales for development and implementation

There are too many questions and unknowns at this stage to determine whether iron fertilisation represents a feasible and practical approach for enhancing ocean carbon sequestration. It would be premature to assign a programme of implementation without further research into application and distribution technologies and long-term environmental impact assessments.

2.9 Verification

Verification of the impact of any iron fertilisation will be particularly difficult bearing in mind the relatively remote location of the HNLC regions. Although the scale and extent of the increase in phytoplankton productivity can be assessed remotely via satellite ocean colour sensors, this is only an indicator of the response in the mixed surface layer and gives no indication of carbon export. Quantification of carbon export to the deep ocean would require an extensive long-term monitoring programme consisting of particle trapping and sediment sampling both in and around the fertilised site (with concurrent comparison with a similar unfertilised site). The technology to achieve this is available, but the verification programme would be particularly challenging both logistically and practically.

2.10 Next steps/further work

The biological and biogeochemical response to iron addition may vary substantially⁴, and the results from the Equatorial Pacific and Southern Ocean experiments cannot be extrapolated to other regions. It is only by performing an iron fertilisation experiment in situ that we are able to determine the actual biological and biogeochemical response. The success of the iron fertilisation experiments in the Pacific and Southern Oceans was achieved by releasing the iron with the tracer sulphur hexafluoride (SF₆) which acted as surrogate for the iron when levels decreased below detection limits (Fig.2). Ecosystem manipulation experiments of this kind are considerably more effective in determining the actual in situ response than extrapolation of oceanographic observations, in vitro laboratory incubations and modelling scenarios.

The way forward for determining the validity of iron fertilisation is the initiation of a long-term continuous observation programme at a fixed site within a HNLC region. The shortcomings of extrapolating ecosystem response dictates that the chosen region should be that identified as the most favourable by modelling and field studies, which would appear to be the convective plume regions in the Southern Ocean.

The programme should focus upon the ecosystem/biogeochemical response within the surficial and deep water column and sediment, with a monitoring programme developed from hydrodynamic considerations including the dispersal of detrital material below the mixed surface layer and the transport of deep water away from the region. The programme should include an air-sea component to monitor production and exchange of the gases described above, with the aim of developing a net biogas GWP weighted budget to determine the integrated impact of iron fertilisation. In addition it is essential that that the biogeochemical and biological status of the water column and sediment are monitored and compared with external unfertilised sites. A programme of this size and scale would require significant funding.

In the IRONEX2 surface water nitrate concentrations only declined to 50% of the initial concentrations, suggesting that addition of iron may not result in full macro-nutrient depletion. Furthermore, modelling work showed that the effect of nutrient depletion was reversible; if nutrient depletion ceased after 50 years following the cessation of iron addition some of the CO₂ captured would return to the atmosphere during the subsequent 50 years^3,20. As most modelling studies are based upon complete utilisation of nutrients it is essential that this is actually achieved in the field before iron fertilisation can be regarded as a viable option. This suggests that further fieldwork on intermediate timescales of approximately 2-3 months is required to establish this, before inception of the long-term monitoring programmes described above

As also noted above iron fertilisation may stimulate nitrogen-fixation¹⁹, and further increase the nutrient and phytoplanktonic productivity status of the oceans. This is also requires further experimental investigation in the field.

3. FERTILISING THE OCEAN WITH MACRO-NUTRIENTS

   3.1 Description

   Macro-nutrients are typically defined as nitrate (+ ammonia), phosphate and silicate. The concentrations of these compounds vary both spatially and temporally in response to ocean circulation. Water upwelled from the deep oceans will contain high levels of macro-nutrients, algal growth can reduce macro-nutrient concentrations to close to zero and is seasonally variable, while inputs are predominantly from the atmosphere and from land runoff. The world's oceans are thought to be predominantly limited by the availability of nitrogen (in the form of nitrate/nitrite and ammonia) although silicate limitation is not uncommon. Unlike the case of iron, fertilising existing HNLC areas with macro-nutrients would make little sense as, by definition, these areas are already replete with these compounds. Nutrients would need to be added to those areas of the oceans where they are currently limiting algal growth either year round or on a seasonal basis only. However, predicting the effect of deliberate ocean fertilisation on atmospheric CO₂ is not simple and, as a consequence, can be predicted incorrectly.

   The role of ocean chemistry is critically important and sometimes neglected by climate engineering enthusiasts. As discussed in the introduction, most of the inorganic carbon in seawater is present as HCO₃^- not CO₂. Although an increase in algal biomass due to nutrient enrichment would remove inorganic carbon from the surface ocean, there is almost an order of magnitude lower effect in the size of the reduction in dissolved CO₂ and hence a relatively small effect on the flux of CO₂ across the air-sea interface. Put more simply, little of the carbon removed from the water by increasing algal growth is immediately replaced by CO₂ from the atmosphere.

   Again, the key aim of any macro-nutrient fertilisation would be to increase either the biological or solubility pump or both. Therefore, for nutrient enrichment to have any net effect on atmospheric CO₂ via the biological pump there also has to be an increase in the carbon export from the surface oceans. Otherwise, the carbon taken up by the algae will be recycled back to CO₂ (and hence HCO₃^-) in the surface oceans with no net effect on atmospheric CO₂. The only way that fertilisation by macro-nutrients can affect the solubility pump is by targeting those areas where surface water is being transported into the deep oceans.

   There are several ways in which the ocean fertilisation by macro-nutrients could be targeted. The most obvious of these, but least investigated, is to add nutrients in order to stimulate phytoplankton growth. Indeed, theoretically it ought to be possible to bias the production of different classes of algae by adding different proportions of nutrients e.g. addition of silicate to an area such as the North Atlantic might be expected to result in more diatom species compared to coccolithophores thereby enhancing ocean uptake of atmospheric CO₂.

   The most heavily researched idea is to that of producing massive floating seaweed farms that would cover large areas of the surface oceans in kelps (brown seaweeds) and in particular the giant Macrocystis species. This is reported on in more detail later. Another variant of this is to grow large floating beds of calcareous algae (Neushul 1991). These are seaweeds containing solid calcium carbonate (CaCO₃). Although superficially, this might be thought to result in an increased CaCO₃ flux to the sediments and hence removal of CO₂ from the surface waters, the idea ignores ocean chemistry. Removal of CO₃^2- to form CaCO₃ actually results in increased CO₂ and hence is not a way of sequestering atmospheric CO₂. This idea is not considered further.

   3.2 Previous and on-going research

   Previous research has predominantly focussed on the potential for producing large kelp farms in the oceans. Much of this effort has been reviewed by Ritschard (1992) while the possible impact of kelp farms on atmospheric CO₂ has been modelled by Orr and Sarmiento (1992). The main conclusions of the latter study are that use of kelp farms to sequester atmospheric CO₂ would be expensive and inefficient. It would require close to the world's current rate of fertiliser production for phosphate, and substantially greater than that for nitrate, to be added to the oceans in order to sequester 0.7GtCyr^-1. In this calculation, it was also assumed that the algae would be harvested (i.e. removed to the land). This does not seem to be feasible. The authors predict that without harvesting, the uptake due to macroalgae was only 0.44 Gt C yr^-1. It is not clear to us how the authors determined the size of any export flux of organic carbon from the surface layer to the deep oceans. We suspect that this is unlikely to be significantly enhanced by the presence of macroalgal beds and that the actual sequestration is likely to be considerably less than 0.44 Gt C yr^-1.

   Little research into the direct use of macro-nutrients for CO₂ sequestration appears to be on-going. A Japanese funded experiment is currently underway with the aim of increasing fish catches and sequestering atmospheric CO₂ (Matuo et al. 1995). Similarly, much of the other on-going research summarised by Omerod and Angel (1998) is primarily aimed at fertilising the oceans in order to increase ocean harvests (typically fish catches). These have the potential to enhance the ocean uptake of atmospheric CO₂ as a 'by-product' of increases in ocean productivity. One of these projects is the Norwegian MARICULT programme whose primary aim is to increase fish production by the application of fertilisers to the open ocean. Omerod and Angel (1998) also report on an experimental programme to try to stimulate a phytoplankton bloom that would fix atmospheric nitrogen and hence enhance growth of biomass at successively higher trophic levels (Markels 1997).

   However, as discussed by Omerod and Angel (1998), the two aims of increasing oceanic uptake of CO₂ and increasing fish production are not necessarily compatible. Maximising the export flux of organic material would benefit the oceanic uptake of CO₂, whereas keeping the organic material in surface waters to maximise grazing would benefit fish production.

   3.3 Potential scale for approach by governments

   It is not clear to us that the main idea behind ocean fertilisation using macro-nutrients, ie the use of kelp farms, is likely to achieve its aims. The modelling work of Orr and Sarmiento (1992) envisages enhanced macroalgal growth across most of the world's oceans and appears to have been as much an intellectual challenge as a serious proposal. It has to be considered that that the amount of energy expended in creating and maintaining the macroalgal farms must be outweighed by the amount of CO₂ sequestered into the deep oceans.

   3.4 Likely environmental impact

The environmental impact of large-scale macro-nutrient fertilisation is likely to be substantial. There are several major environmental concerns associated with ocean fertilisation strategies in general and with macroalgal beds in particular. These are discussed below:

1. The reduction in biodiversity in the surface water of areas of ocean where primary production has been increased. Any fertilisation scheme is bound by its very nature to affect the species biodiversity in the surface layer. As discussed in the introduction, the marine plankton in these areas of the ocean are already responsible for internally cycling 25 times the amount of anthropogenic CO₂ that is currently entering the oceans every year. Similarly, there is already a natural exchange of CO₂ across the air-sea interface that is 45 times the ocean uptake of anthropogenic CO₂. A small change to this natural cycle would have a disproportionately huge effect on the net air-ocean CO₂ flux. We know too little about how the plankton community in the surface ocean would respond to fertilisation and about the feedback mechanisms (see below) that might result from population changes.
2. There is a general concern that the marine community structure will be changed with unknown consequences throughout the food chain.
3. There is a danger of creating oxygen-deficient areas in deeper and adjacent waters resulting from increased microbial respiration. The study of Orr and Sarmiento (1992) suggests that this is likely if their figure of only 0.44 Gt C yr^-1 is sequestered by macroalgal farms. Not only does this have important implications for the use of the ocean as a food resource, there is also the potential for increased fluxes of CH₄ and N₂O from the ocean to the atmosphere (Fuhrman and Capone 1991). Both of these gases are greenhouse gases (see introduction) and might well counteract any benefits accruing from CO₂ sequestration.
4. There is the likelihood of negative feedback mechanisms via the production of other biogenic gases that may affect atmospheric chemistry and climate. For example, macroalgal are known to release several iodine, bromine and chlorine containing compounds such as chloroform, methyl bromide, methyl iodide, dibromomethane, dichlorobromomethane and bromoform (e.g. Nightingale et al. 1995). These gases have been implicated in ozone depletion and are important in atmospheric chemistry. High levels of dimethylsulphoniopropionate (DMSP) have been observed in macroalgal tissues. DMSP is the precursor of dimethyl sulphide which is thought to have a large influence on climate via the formation of cloud condensation nuclei in the atmosphere (Charlson et al. 1987). In addition, epiphyton associated with macroalgae have been identified as sources of nitrous oxide (Law et al. 1993), another greenhouse gas that also represents a sink for stratospheric ozone.
5. Clearly, fertilisation could only be considered at a level at which there was no significant environmental detriment. The question of whether there is indeed an environmentally acceptable level of fertilisation needs to be addressed

3.5 Other effects

Omerod and Angel (1998) argue that the concept of ocean fertilisation becomes more attractive if the increased oceanic uptake of CO₂ from the atmosphere is a benefit supplementary to other goals. This would be the case if fertilisation could be wholly, or partly, justified in terms of an increased fish catch or by growth of macroalgae as a raw material for a biofuel or other chemicals. Nutrient additions to lakes on Vancouver Island, Canada resulted in increased catches of Sockeye salmon (Le Brasseur et al., 1979). As discussed by Omerod and Angel (1998), marine macroalgae are already used in food processing industries in China, Japan and the Philippines. Specific polysaccharides are extracted from macroalgae for use in the biotechnology and pharmaceutical industries; macroalgal biomass is converted into fertiliser and soil conditioning products.

However, there is some doubt as to whether the two aims of enhancing of macroalgal growth to sequester CO₂ and increasing fish production, or use of macroalgae on land, are mutually compatible. Maximising the export flux of organic material would benefit the oceanic uptake of CO₂, whereas keeping the organic material in surface waters to maximise grazing would benefit fish production. Also, in order to sequester CO₂ into the ocean, the optimum location for macroalgae would be in the deep ocean not in areas where macroalgae could be economically harvested. If, as has been suggested, the macroalgae were used as a fuel source on land then it could result in an increase in atmospheric CO₂ concentration depending on how the harvest was utilised.

3.6 Economic costs

The use of kelp farms to sequester atmospheric CO₂ is very difficult to cost with any degree of confidence. As a result, we only give a range of costs below. It is inevitable given our present lack of knowledge that some costs can only be very approximately estimated at (e.g. costs of delivery of nutrients to the oceans). Other costs are indirect (e.g. possible production of other greenhouse gases by the macroalgae), but ought to be included in future estimates and there are bound to be others that are presently unforeseen.

The modelled results of Orr and Sarmiento (1992) show that the use of kelp farms to sequester atmospheric CO₂ would be expensive and inefficient. It would require close to the world's current rate of fertiliser production for phosphate, and substantially greater than that for nitrate, to be added to the oceans in order to sequester 0.44GtCyr^-1. Additionally, the emissions of CO₂ from manufacture of nitrogen fertiliser would be a significant fraction of the amount of CO₂ drawn-down.

Omerod and Angel (1998) have assessed the likely cost of macro-nutrient fertilisation based on the work of Jones and Otaegui (1997) and suggest doubling the latter authors estimates to between £30 and £50/tC. Ritschard (1992) estimated the cost of open ocean kelp farms to be about £120/tC. It is not clear to us whether these costs are actually per tonne of C as atmospheric CO₂ converted to organic carbon or more importantly per tonne of C sequestered i.e. exported to the deep ocean or harvested and returned and locked on land. A more thorough costing of macro-nutrient fertilisation should be a priority.

3.7 Practicality

The logistics of fertilising large areas of the ocean with macro-nutrients are likely to be enormous and we do not attempt to comment on them further.

3.8 Likely time scales for development and implementation

The concepts of ocean fertilisation discussed are all relatively novel and there are considerable scientific uncertainties associated with them. These need to be resolved before any plans are made for use of this option. This will involve a significant amount of basic research.

3.9 Verification

Methods of verifying the amount of carbon sequestered in the ocean need to be developed and this would be a great challenge. It would appear likely that this would require a combination of on-site measurement of relevant parameters (CO₂ speciation, export flux, and environmentally critical parameters) and ocean models with which to place the filed measurements in a wider regional and global context.

3.10 Next steps/further work

Of critical importance with respect to the use of ocean fertilisation with macro-nutrients is to investigate whether there is an environmentally acceptable level to which the oceans can be artificially stimulated to sequester anthropogenic CO₂. Important in providing an answer to this are the following.

We currently know too little about the extremely large natural ocean carbon cycle to predict with any degree of confidence the long term effects of a deliberate perturbation (i.e. ocean fertilisation) on it.

Research programmes would be required into the effects on species biodiversity and the marine community of macro-nutrient enrichment. These programmes of research into ocean productivity should include the effects on carbon export from the surface oceans as this is a critical factor as to whether any enhanced sequestration of carbon is achieved.

Research would also be required as to whether there would be other less desirable and harder to predict effects of ocean fertilisation. One obvious candidate is the possibility of reduced oxygen concentrations in the water column. Other potential feedback mechanisms eg the production of other biogenic gases that may be important in global climate, require urgent attention.

Large field programmes would be required that would need to be developed in conjunction with ocean carbon cycle modellers so that results from field sites could be placed in a wider context.

REFERENCES FOR IRON FERTILISATION

¹ Frost, B. W. 1991. Limnol.Oceanogr. 36 1616-1630

² Smith, W. O., & Nelson, D. M., 1990. Limnol. Oceanogr 35, 809-821

³ Martin, J.H. and S.E. Fitzwater, Nature, 1988. 331:341-343; Martin, J.H. et al, Nature, 1990. 345:156-158, Boyd, P.W et al. Mar. Ecol. Prog. Ser. 105:21-32.

⁴ Martin J.H. et al, 1994, Nature, 371:123-129; Watson et al, 1994, Nature, 371:143-145; Coale, K.H., et al., 1996, Nature, 383:495-501; Cooper, D.J. et al, 1996, Nature, 383: 511-513; Turner, S.M., et al., 1996, Nature,. 383:513-517; Behrenfeld, M.J., et al., 1996. Nature, 383: 508-511.

⁵ Siegenthaler, U. and J.L. Sarmiento, 1993, Nature, 365:119-125.

⁶ Knox, F., and M. McElroy, 1984. J. Geophys. Res 89: 4629-4637; Sarmiento, J. L., and J. R. Toggweiler, 1984, Nature 308: 621-624.

⁷ Sarmiento, J.L. and J.C. Orr. Limnol. Oceanogr., 36:1928-1950; Peng T.-H. and Broecker, W.S., 1991, Nature, 349: 227-229: Joos et al, 1991, Nature, 349: 772-775; Kurz, K.D. and E. Maier-Reimer, 1993. Glob. Biogeochem. Cycles, 7:229-244.

⁸ Martin, J. 1990. Palaeoceanog. 5: 1-13.

⁹ Pedersen, T.F. et al, 1991. Palaeoceanog 6: 657-677.

¹⁰ Watson, A.J. 1997. Nature 385: 587-588.

¹¹ Charleson, R.J. et al, 1987. Nature 326:655-661.

¹² Butler et al, 1989.J. Geophys. Res.94:18265-18271, Law, C.S. and N.J.P. Owens, 1990, Nature, 346: 826-828, Owens, N.J.P. et al, 1991, Nature, 354: 93-296.

¹³ Fuhrman, J.A. and D.G.Capone, 1991. Limnol.Oceanog. 36: 1951-1959.

¹⁴ Houghton, J.T. et al, 1996. Climate Change 1995.IPCC. Cambridge Univ. Press

¹⁵ Martin, J.H., Fitzwater, S.E. ,and R.M. Gordon. Global Biogeochemical cycles 4:5-12.

¹⁶ Peng, T-H and W.S. Broecker, Nature, 349: 227-229.

¹⁷ Boyd, P.W. et al., Global Biogeochemical Cycles 12:429:443.

¹⁸ Boyd, P.W., Watson, A.J. and C.S. Law. U.S. JGOFS Newsletter, February 1999.

¹⁹ Falkowski, P.G. Nature 387: 272-275.

²⁰Ormerod, W and M. Angel. Ocean fertilisation as a CO₂ sequestration option. IEA Greenhouse Gas R&D Programme Report

²¹Ritschard, R.L. 1992. Water, Soil and Air Poll. 64: 289-303

REFERENCES FOR MACRO-NUTRIENT FERTILISATION

Charlson R.J. et al. (1987), Nature 326 655-661.

Fuhrman J.A. and D.G. Capone (1991) Limnol and Oceanogr. 36 1951-1959.

Houghton J.T. et al., (1996), Climate Change 1995, The Science of Climate Change, Cambridge University Press, ISBN 0 521 56436 0.

Jones I.S.F. and D. Otaegui (1997), Energy Convers. Mgmt. 38 (suppl) 367-372.

Law C.S., A.P. Rees and N.J.P. Owens (1993), Limnol and Oceanogr. 38 43-441.

Le Brasseur et al., (1979) in Omerod W. and M. Angel, (1998), IEA Greenhouse Gas R&D Programme, ISBN 1 898373 11 6.

Matuo E. et al., (1995), Energy Conversion and Management 36, 907-910.

Markels M., (1997) Second International Conference on Open Ocean Aquaculture.

Neushul M. (1991), EPRI Report (EAK7-401), Electric Power research Inst. Palo Alto, CA, USA.

Nightingale P.D., G. Malin and P.S. Liss (1995), Limnol and Oceanogr. 40 680-689.

Omerod W. and M. Angel, (1998), IEA Greenhouse Gas R&D Programme, ISBN 1 898373 11 6.

Orr J.C. and J.L. Sarmiento, (1992), Water Air and Soil Pollution 64 405-421.

Ritschard R.L., (1992), Water Air and Soil Pollution 64 289-303.

---

Previous | Home Page | Next

---

Updated 16 July 1999

---

Go to DETR Sustainable Development Index

---

Go to Cabinet Office Home Page