...a rotary pump which circulates material in the microplankton food web of the euphotic zone, an Archimedian pump by which the flux of fecal and aggregated material occurs continually under gravity, and a reciprocating pump by which diel migrants actively carry material down at dawn, to rise again at dusk to feed.

To these we must certainly add a diffusion pump by which dissolved and non-sinking particulate organic matter accumulate in the photic zone and are mixed downward (Anderson, 1993).

To date, most sediment trap studies in the Southern Ocean have shown the biogenic particle flux to be highly seasonal, usually restricted to a brief summer period, with integrated annual fluxes that are very low in comparison with those in other oceans (e.g., Fischer et al., 1988; Honjo, 1990). Those fluxes typically represent only about 10% of the annual primary productivity, even though primary productivity is very low by global ocean standards (see above). Thus the Archimedian variant of the biological pump (sensu Longhurst, 1991) appears to operate slowly and sporadically in the Southern Ocean. Mid-summer fluxes in spatially restricted areas of high productivity such as the Ross Sea can be enormous (e.g., DeMaster et al., 1992), but the small extent of those areas makes their overall contribution to organic matter export minimal.

Studies of nitrogen uptake by phytoplankton in the Southern Ocean suggest that about 50% of the primary production is supported by nitrate, irrespective of the total primary productivity of a region (Smith and Sakshaug, 1990; Nelson, 1992). Because nitrate is supplied from beneath the euphotic zone and other nitrogenous nutrients (primarily ammonium and urea) are supplied by in situ recycling within the euphotic zone (Dugdale and Goering, 1967), the ratio of nitrate uptake to total nitrogen uptake by phytoplankton (the f ratio; Eppley and Peterson, 1979) can be taken as a measure of the fraction of primary production that is potentially available for export to depth. Thus about 50% of the organic matter produced by phytoplankton appears to be potentially available for export in the Southern Ocean, over a wide range of regional primary productivity levels. This finding contrasts sharply with observations in tropical and subtropical systems, where the general pattern is that only areas of high annual primary productivity have f ratios as high as 0.5, with values as low as 0.05 - 0.1 in oligotrophic surface waters (e.g., Eppley and Peterson, 1979).

If the existing data sets on gravitational particle flux and f ratios in the Southern Ocean are both correct in their overall magnitudes, then one of two things must be true. Either:

1. there is considerable vertical export of organic matter from the euphotic zone of the Southern Ocean via pathways other than gravitational settling (e.g., active migration of grazing organisms or downward mixing of dissolved organic material and non-sinking particles), or

2. nitrogen-based f ratios provide large overestimates of the production of potentially exportable organic matter in the Southern Ocean.

Implications for field measurements: While it will be important to measure the gravitational flux of biogenic particulate matter with sediment traps, and to do everything possible to assure that the sediment trap data provide an accurate measure of the gravitational flux term, it will be of at least equal importance to obtain data that permit estimates of non-gravitational fluxes. Studies of grazing rates, especially those of organisms large enough to undertake significant diel vertical migrations, will be necessary to quantify the "reciprocating pump" as a transport mechanism for organic matter. Similarly, studies of the rate of turbulent mixing near the base of the euphotic zone and detailed measurements of the vertical gradients in dissolved organic matter and suspended (non-sinking) biogenic particulate matter will be necessary to evaluate the downward mixing "diffusion pump" term. Both of these non-gravitational mechanisms may be sharply discontinuous in time (e.g., they may be driven by major events such as storms or the passage of krill swarms). Sampling strategies should take that strong likelihood into account.

Implications for modeling: It would be very useful to determine whether a nitrogen-based f ratio is a valid measure of the potential fraction of primary productivity that can be exported to depth in a system where nitrate is not depleted to potentially limiting concentrations. Qualitative arguments for both alternatives can be presented: On one hand, the strong positive correlation between the nitrogen-based f ratio and total primary productivity in tropical and subtropical systems may be based largely on differences in nitrate availability in the surface layer. This would predict that multiplying f by productivity may be a poor predictor of organic matter export in nitrate-replete systems such as the Southern Ocean. On the other hand, the overall character of the nitrogen cycle in the Southern Ocean is similar to that in other areas; nitrate is regenerated at depth and advected or mixed upward, while ammonium and urea are regenerated within the surface layer. This would predict that the product of f and productivity is closely related to export even in systems where no nutrient is limiting to the phytoplankton. These alternatives may be addressable in quantitative terms by modeling, which if done in advance of field programs may help identify those field data that are most crucial to resolving the question.