The net primary production of marine ecosystems is roughly the same as for terrestrial ecosystems (50GtC/yr for marine ecosystems and 60GtC/yr for terrestrial ecosystems), and there are opportunities to increase the net carbon flow into the marine biosphere. There are fundamental differences between the two systems, however, as the marine biosphere does not include large stores of carbon in the living and dead biomass. There are some 3 GtC in marine biota versus nearly 2500GtC in terrestrial vegetation and soils (Table 4.1). The key to increasing the carbon stocks in ocean ecosystems is thus to move carbon through the small reservoir of the marine biota to the larger reservoirs of dissolved inorganic carbon (the biological pump) in ways that will isolate the carbon and prevent its prompt return to the atmosphere. The biological pump serves to move carbon from the atmosphere to the deep oceans, as organisms take up CO2 by photosynthesis in the surface ocean, and release the carbon when the organic material sinks and is oxidized at depth.
Several researchers have suggested that ocean productivity in major geographical regions is limited by the availability of primary or micronutrients, and that productivity could be increased substantially by artificially providing the limiting nutrients. This might involve providing nitrogen or phosphorus in large quantities, but the quantities to be supplied would be much smaller if growth were limited by a micronutrient. In particular, there is evidence that in large areas of the Southern Ocean productivity is limited by availability of the micronutrient iron. Martin (1990, 1991) suggested that the ocean could be stimulated to take up additional CO2 from the atmosphere by providing additional iron, and that 300,000 tonnes of iron could result in the removal of 0.8GtC from the atmosphere. Other analyses have suggested that the effect may be more limited. Peng and Broecker (1991) examined the dynamic aspects of this proposal and concluded that, even if the iron hypothesis was completely correct, the dynamic issues of mixing the excess carbon into the deep ocean would limit the magnitude of the impact on the atmosphere. Joos et al. (1991) reported on a similar model experiment and found the ocean dynamics to be less important, the time path of anthropogenic CO2 emissions to be very important, and the maximum potential effect of iron fertilization to be somewhat greater than reported by Peng and Broecker (1991).
Some of the concepts of iron fertilization have now been tested with 2 small-scale experiments in the equatorial Pacific Ocean. In experiment IronEX 1 (November, 1993) 480 kg of iron were added over 24 hours to a 64 km2 area of the equatorial Pacific. In IronEX 2 (May/June, 1995) a similar 450 kg of iron (as acidic iron sulphate) were added over a 72 km2 area, but the addition occurred in 3 doses over a period of one week.
The IronEx 1 experiment showed unequivocally that there was a biological response to the addition of iron. However, although plant biomass doubled and phytoplankton production increased fourfold, the decrease in CO2 fugacity (in effect the partial pressure of CO2 decreased by 10 micro atm) was only about a tenth of that expected (Martin et al., 1994; Watson et al., 1994; Wells, 1994). In the IronEX 2 experiment the abundance and growth rate of phytoplankton increased dramatically (by greater than 20 and twice, respectively), nitrate decreased by half, and CO2 concentrations were significantly reduced (the fugacity of CO2 was down 90matm on day 9). Within a week of the last fertilization, however, the phytoplankton bloom had waned, the iron concentration had decreased below ambient, and there was no sign that the iron was retained and recycled in the surface waters (Monastersky, 1995; Coale et al., 1996; Cooper et al., 1996; Frost, 1996).
These two experiments have demonstrated that week-long, sustained additions of iron to nutrient-rich, but iron-poor, regions of the ocean can produce massive phytoplankton blooms and large drawdowns of CO2 and nutrients. While the results of these two experiments cannot be uncritically extrapolated, they suggest a very important role for iron in the cycling of carbon (Cooper et al., 1996). The consequences of larger, longer-term introductions of iron remain uncertain. Concerns that have been expressed relate to the differential impact on different algal species, the impact on concentrations of dimethyl sulphide in surface waters, and the potential for creating anoxic regions at depth (Coale et al., 1996; Frost, 1996; Turner et al., 1996). There is much to be learned of the ecological consequences of large-scale fertilization of the ocean.
Jones and Young (1998) suggest that the addition of reactive nitrogen in appropriate areas, perhaps in conjunction with trace nutrients, would increase production of phytoplankton and could both increase CO2 uptake and provide a sustainable fishery with greater yield than at present.
Chemical buffering of the oceans to decreases in pH associated with uptake of CO2 leads to an increase in dissolved inorganic carbon that does not rely on alteration of the biological pump. Buffering of the oceans is enhanced by dissolution of alkaline minerals. Dissolution of alkaline materials in ocean sediments with rising pH occurs in nature, but does so on a time-scale of thousands of years or more (Archer et al., 1997). Intentional dissolution of mined minerals has been considered, but the quantity (in moles) of dissolved minerals would be comparable to the quantity of additional carbon taken up by the oceans (Kheshgi, 1995).
Stallard (1998) has shown that human modifications of the earths surface may be leading to increased carbon stocks in lakes, water reservoirs, paddy fields, and flood plains as deposited sediments. Burial of 0.6 to 1.5GtC/yr may be possible theoretically. Although Stallard (1998) does not suggest intentional manipulation for the purpose of increasing carbon stocks, it is clear that human activities are likely leading to carbon sequestration in these environments already, that there are opportunities to manage carbon via these processes, and that the rate of carbon sequestration could be either increased or decreased as a consequence of human decisions on how to manage the hydrological cycle and sedimentation processes.
The term geo-engineering has been used to characterize large-scale, deliberate manipulations of earth environments (NAS, 1992; Marland, 1996; Flannery et al., 1997). Keith (2001) emphasizes that it is the deliberateness that distinguishes geo-engineering from other large-scale, human impacts on the global environment; impacts such as those that result from large-scale agriculture, global forestry activities, or fossil fuel combustion. Management of the biosphere, as discussed in this chapter, has sometimes been included under the heading of geo-engineering (e.g., NAS, 1992) although the original usage of the term geo-engineering was in reference to a proposal to collect CO2 at power plants and inject it into deep ocean waters (Marchetti, 1976). The concept of geo-engineering also includes the possibility of engineering the earths climate system by large-scale manipulation of the global energy balance. It has been estimated, for example, that the mean effect on the earth surface energy balance from a doubling of CO2 could be offset by an increase of 1.5% to 2% in the earths albedo, i.e. by reflecting additional incoming solar radiation back into space. Because these later concepts offer a potential approach for mitigating changes in the global climate, and because they are treated nowhere else in this volume, these additional geo-engineering concepts are introduced briefly here.
Summaries by Early (1989), NAS (1992), and Flannery et al. (1997) consider a variety of ways by which the albedo of the earth might be increased to try to compensate for an increase in the concentration of infrared absorbing gases in the atmosphere (see also Dickinson, 1996). The possibilities include atmospheric aerosols, reflective balloons, and space mirrors. Most recently, work by Teller et al. (1997) has re-examined the possibility of optical scattering, either in space or in the stratosphere, to alter the earths albedo and thus to modulate climate. The latter work captures the essence of the concept and is summarized briefly here to provide an example of what is envisioned. In agreement with the 1992 NAS study, Teller et al. (1997) found that ~107t of dielectric aerosols of ~100 nm diameter would be sufficient to increase the albedo of the earth by ~1%. They showed that the required mass of a system based on alumina particles would be similar to that of a system based on sulphuric acid aerosol, but the alumina particles offer different environmental impact. In addition, Teller et al. (1997) demonstrate that use of metallic or optically resonant scatterers can, in principle, greatly reduce the required total mass of scattering particles required. Two configurations of metal scatterers that were analyzed in detail are mesh microstructures and micro-balloons. Conductive metal mesh is the most mass-efficient configuration. The thickness of the mesh wires is determined by the skin-depth of optical radiation in the metal, about 20 nm, and the spacing of wires is determined by the wavelength of scattered light, about 300nm. In principle, only ~105t of such mesh structures are required to achieve the benchmark 1% increase in albedo. The proposed metal balloons have diameters of ~4 mm and a skin thickness of ~20nm. They are hydrogen filled and are designed to float at altitudes of ~25km. The total mass of the balloon system would be ~106t. Because of the much longer stratospheric residence time of the balloon system, the required mass flux (e.g., tonnes replaced per year) to sustain the two systems would be comparable. Finally, Teller et al. (1997) show that either system, if fabricated in aluminium, can be designed to have long stratospheric lifetimes yet oxidize rapidly in the troposphere, ensuring that few particles are deposited on the surface.
One of the perennial concerns about possibilities for modifying the earths radiation balance has been that even if these methods could compensate for increased GHGs in the global and annual mean, they might have very different spatial and temporal effects and impact the regional and seasonal climates in a very different way than GHGs. Recent analyses using the CCM3 climate model (Govindasamy and Caldeira, 2000) suggest, however, that a 1.7% decrease in solar luminosity would closely counterbalance a doubling of CO2 at the regional and seasonal scale (in addition to that at the global and annual scale) despite differences in radiative forcing patterns.
It is unclear whether the cost of these novel scattering systems would be less than that of the older proposals, as is claimed by Teller et al. (1997), because although the system mass would be less, the scatterers may be much more costly to fabricate. However, it is unlikely that cost would play an important role in the decision to deploy such a system. Even if we accept the higher cost estimates of the NAS (1992) study, the cost may be very small compared to the cost of other mitigation options (Schelling, 1996). It is likely that issues of risk, politics (Bodansky, 1996), and environmental ethics (Jamieson, 1996) will prove to be the decisive factors in real choices about implementation. The importance of the novel scattering systems is not in minimizing cost, but in their potential to minimize risk. Two of the key problems with earlier proposals were the potential impact on atmospheric chemistry, and the change in the ratio of direct to diffuse solar radiation, and the associated whitening of the visual appearance of the sky. The proposals of Teller el al. (1997) suggest that the location, scattering properties, and chemical reactivity of the scatterers could, in principle, be tuned to minimize both of these impacts. Nonetheless, most papers on geo-engineering contain expressions of concern about unexpected environmental impacts, our lack of complete understanding of the systems involved, and concerns with the legal and ethical implications (NAS, 1992; Flannery et al., 1997; Keith, 2000). Unlike other strategies, geo-engineering addresses the symptoms rather than the causes of climate change.
Other reports in this collection