6.04 The C. L. de la Rocha University of Cambridge, UK

NOMENCLATURE 84 6.04.1 INTRODUCTION 84 6.04.2 DESCRIPTION OF THE BIOLOGICAL PUMP 84 6.04.2.1 and Uptake 86 6.04.2.1.1 Levels of 87 6.04.2.1.2 Patterns in time and space 87 6.04.2.1.3 Nutrient limitation 88 6.04.2.2 and Sinking 88 6.04.2.2.1 88 6.04.2.2.2 Aggregation and exopolymers 89 6.04.2.2.3 Sinking 89 6.04.2.3 Particle and Repackaging 90 6.04.2.3.1 grazing 90 6.04.2.3.2 Bacterial hydrolysis 90 6.04.2.3.3 Geochemistry of decomposition 90 6.04.2.4 and Burial 91 6.04.2.5 Dissolved 91 6.04.2.6 New, Export, and Regenerated Production 92 6.04.3 IMPACT OF THE BIOLOGICAL PUMP ON GEOCHEMICAL CYCLING 92 6.04.3.1 Macronutrients 92 6.04.3.1.1 92 6.04.3.1.2 92 6.04.3.1.3 93 6.04.3.1.4 94 6.04.3.2 Trace Elements 96 6.04.3.2.1 96 6.04.3.2.2 96 6.04.3.2.3 Cadmium 98 6.04.3.2.4 98 6.04.4 QUANTIFYING THE BIOLOGICAL PUMP 99 6.04.4.1 Measurement of New Production 99 6.04.4.2 Measurement of Particle Flux 100 6.04.4.2.1 traps 100 6.04.4.2.2 Particle-reactive nuclides 101 6.04.4.2.3 Oxygen utilization rates 101 6.04.5 THE EFFICIENCY OF THE BIOLOGICAL PUMP 102 6.04.5.1 Altering the Efficiency of the Biological Pump 102 6.04.5.1.1 In HNLC areas 102 6.04.5.1.2 Through changes in community composition 104 6.04.5.1.3 By varying the C : N : P ratios of sinking material 104 6.04.5.1.4 By enhancing particle transport 105 6.04.6 THE BIOLOGICAL PUMP IN THE IMMEDIATE FUTURE 105 6.04.6.1 Response to Increased CO2 105 6.04.6.2 Response to Agricultural Runoff 106 6.04.6.2.1 Shift towards silicon limitation 106 6.04.6.2.2 Shifts in export production and deep C : N : P 106 6.04.6.3 via Ocean Fertilization and the Biological Pump 106 ACKNOWLEDGMENTS 107 REFERENCES 107

83 84 The Biological Pump NOMENCLATURE One side-effect of the biological pump is that CO is shunted from the surface ocean and into * 2 H2CO3 dissolved CO2 þ H2CO3 (mM) the deep , thus lowering the amount in the J flux of organic C to depth Corg atmosphere. For many years it has been (g C m22 yr21) recognized that pre-Industrial CO2 levels in the z depth (m) atmosphere were about one-third of what they PP primary production (g C m22 yr21) 2 21 would be in the absence of a biological pump D diffusivity of CO2 in seawater (m s ) (Broecker, 1982). It is also known that the r radius of cell (mm) biological pump is not operating at its full Ce extracellular CO2 concentration (mM) capacity. In so-called “high-nutrient, low- Ci intracellular CO2 concentration (mM) 0 21 chlorophyll” (HNLC) areas of the ocean, a k rate constant for HCO3 ! CO2 (s ) considerable portion of the supplied F flux of CO to cell surface (mmol s21) CO2 2 to the surface waters is not utilized in support of Pt residence time (yr) primary production, most likely due to the CO2 total CO2 (mM) limitation of phytoplankton growth by an inadequate supply of trace elements (e.g., Martin and Fitzwater, 1988). The possibility that the 6.04.1 INTRODUCTION biological pump in HNLC areas might be Despite having residence times (t) that exceed stimulated by massive additions of iron both the ,1,000 yr mixing time of the ocean (Broecker artificially as a means of removing anthropogenic and Peng, 1982), many dissolved constituents of CO2 from the atmosphere and naturally as a seawater have distributions that vary with depth cause for the lower glacial atmospheric CO2 and from place to place. For instance, silicic levels (Martin, 1990) is the focus of much acid (t ¼ 1.5 £ 104 yr), (t ¼ 3,000 yr), research and debate (e.g., Martin et al., 1994; (t ¼ (1–5) £ 104 yr), and dissolved Coale et al., 1996; Boyd et al., 2000; Watson inorganic carbon (DIC; t ¼ 8.3 £ 104 yr) are et al., 2000). generally present in low concentrations in surface Although the biological pump is most popu- waters and at much higher concentrations below larly known for its impact on the cycling of the thermocline (Figure 1). Additionally, their carbon and major nutrients, it also has profound concentrations are higher in older deep waters impacts on the geochemistry of many other than they are in the younger waters of the deep elements and compounds. The biological pump sea (Figure 2). This is the general distribution heavily influences the cycling, concentrations, exhibited by elements and compounds taking and residence times of trace elements—such as part in biological processes in the ocean and cadmium, germanium, zinc, nickel, iron, arsenic, is generally referred to as a “nutrient-type” selenium—through their incorporation into distribution. organic matter and biominerals (Bruland, 1980; Both the lateral and vertical gradients in the Azam and Volcani, 1981; Elderfield and concentrations of nutrients result from “the Rickaby, 2000). Scavenging by sinking biogenic biological pump” (Figure 3). Dissolved inorganic particles and precipitation of materials in the 2 32 microenvironment of organic aggregates and materials (e.g., CO2,NO3 ,PO4 , Si(OH)4) are fixed into particulate organic matter (carbo- fecal pellets plays a large role in the marine hydrates, , ) and biominerals (silica geochemistry of elements such as barium, and ) by phytoplankton in thorium, protactinium, beryllium, rare earth surface waters. Some of these particles are elements (REEs), and yttrium (Dehairs et al., subsequently transported, by sinking, into the 1980; Anderson et al.,1990; Buesseler deep. The bulk of the organic material and et al., 1992; Kumar et al., 1993; Zhang and Nozaki, 1996). Even major elements in seawater biominerals decomposes in the upper ocean via 2þ 2þ dissolution, zooplankton grazing, and microbial such as Ca and Sr display slight surface hydrolysis, but a significant supply of material depletions (Broecker and Peng, 1982; de does survive to reach the and . Villiers, 1999) as a result of the biological Thus, just as biological uptake removes certain pump, despite their long respective oceanic dissolved inorganic materials in surface waters, residence times of 1 Myr and 5 Myr (Broecker the decomposition of sinking biogenic particles and Peng, 1982; Elderfield and Schultz, 1996). provides a source of dissolved inorganic material to deeper waters. Thus, deeper waters contain higher concentrations of biologically utilized 6.04.2 DESCRIPTION OF THE BIOLOGICAL materials than do surface waters. Older deeper PUMP waters contain higher concentrations of bums compared to newly formed deep waters or surface The biological pump can be sectioned into waters. several major steps: the production of organic Description of the Biological Pump 85

P Figure 1 Depth profiles of: (a) CO2, (b) dissolved CO2, (c) silicic acid, (d) nitrate, and (e) phosphate from the Indian Ocean (278 40 S, 568 580 E; GEOSECS Station 427) (source Weiss et al., 1983).

matter and biominerals in surface waters, the . In fact, most of the primary production sinking of these particles to the deep, and the formed will be recycled within the upper hundred decomposition of the settling (or settled) particles. few meters of the (Martin et al., In general, phytoplankton in surface waters take up 1987). Some portion of the primary production DIC and nutrients. Carbon is fixed into organic will, however, be exported to deeper waters or material via photosynthesis and, together with even to the sediments before decomposition and nitrogen, phosphorus, and trace elements, form the may escape remineralization entirely and remain , lipids, and proteins, which all in the sedimentary reservoir. comprise bulk organic matter. Once formed, this It is worth taking a closer look at the various organic matter faces the immediate possibility of steps in the biological pump (Figure 3). Rates, decomposition back to CO2, phosphate, ammonia, overall amounts, and the distribution and char- and other dissolved nutrients through consumption acter of materials produced, transported, and by herbivorous zooplankton and degradation by decomposed vary wildly within the ocean. 86 The Biological Pump

Figure 2 Nitrate concentrations along the great ocean conveyor at 2,000 m depth (source Levitus et al., 1994,by way of the LDEO/IRI Data Library).

Figure 3 Diagram of the biological pump (after OCTET workshop report).

6.04.2.1 Photosynthesis and Nutrient Uptake The first stable product of carbon fixation by the enzyme, ribulose bisphosphate carboxylase In the initial step of the biological pump, (Rubisco), is glyceraldehyde 3-phosphate, a 3-C phytoplankton in sunlit surface waters convert sugar. This 3-C sugar is fed into biosynthetic CO2 into organic matter via photosynthesis: pathways and forms the basis for all organic compounds produced by photosynthetic - CO2 þ H2O þ light –! CH2O þ O2 ð1Þ isms. Fixed carbon and major and trace elements Description of the Biological Pump 87 such as hydrogen, nitrogen, phosphorus, calcium, within cells and are important for silicon, iron, zinc, cadmium, magnesium, iodine, and the maintenance of charge balance (e.g., selenium, and molybdenum are used for the Fagerbakke et al., 1999). synthesis of carbohydrates, lipids, proteins, bio- A wide variety of may be adsorbed onto the , amino acids, enzymes, DNA, and other surfaces of biogenic particles. The removal and necessary biochemicals. deposition of particle-reactive elements such as Besides carbon, the two main components of thorium (Buesseler et al., 1992) and protactinium phytoplankton organic matter are nitrogen and (Kumar et al., 1993) have been shown to correlate phosphorus, in the average molar proportion of with the primary production of particles in the 106C:16N:1P (Redfield, 1934, 1958) known as the ocean. Additionally, thorium has been shown to Redfield ratio. In addition, , by virtue of complex with colloidal, surface-reactive polysac- depositing (amorphous, hydrated silica) in charides (Quigley et al., 2002). their , have an average C:Si ratio of 8 (Brzezinski, 1985), although this ratio may vary from at least 3 to 40 depending on the conditions 6.04.2.1.1 Levels of primary production of light, temperature, and nutrient availability (Harrison et al., 1977; Brzezinski, 1985). Cocco- The amount of primary production carried out in lithophorids produce scales () made of the each year has been estimated from ocean color satellite data and shipboard 14Cincubations CaCO3 and contain 20–100 mmol of CaCO3 per 22 mol of organic carbon (Paasche, 1999). to be 140 g C m for a total of 50–60 Pg C Phytoplankton particulate matter (organic and (4–5 Pmol C) fixed in the surface ocean each year biomineralized) contains many trace elements. (Shuskina, 1985; Martin et al.,1987; Field et al., 1998). This represents roughly half of the global The most abundant are magnesium, cadmium, annual 105 Pg C fixed each year (Field et al.,1998), iron, calcium, barium, copper, nickel, zinc, and despite the fact that marine phytoplankton comprise aluminum (Table 1), which are important con- less than 1% of the total photosynthetic on stituents of enzymes, pigments, and structural Earth. Extrapolation from Redfield ratios suggests materials. Carbonic anhydrase requires zinc or the incorporation 0.6–0.8 Pmol N, 40–50 Tmol P cadmium (Price and Morel, 1990; Lane and into biogenic particles each year in association with Morel, 2000), nitrate reductase requires iron marine primary production. From the proportion of (Geider and LaRoche, 1994), and chlorophyll primary production carried out by diatoms and the contains magnesium. Additionally, elements such average Si:C ratio of diatoms, silica production rates as sodium, magnesium, phosphorus, chlorine, may be calculated to be 200–280 Tmol Si yr21 , and calcium may be present as ions (Nelson et al.,1995; Tre´guer et al.,1995).

Table 1 Elemental composition of marine phytoplankton from cultures and tows. 6.04.2.1.2 Patterns in time and space Element Element : C ratio References Rates of primary in (mol : mol) regions of the ocean outpace those of non-upwelling coastal regions, which in turn are N 0.15 b Si (diatoms only) 0.13 c greater than rates in the oligotrophic open ocean P 0.009 b (Figure 4; Ryther, 1969; Martin et al., 1987). Ca 0.03 d,e Open-ocean primary production levels are 21 Fe 2.3 £ 1026 –1.8 £ 1023 d, e, f ,130 g C m22 yr , whereas in nonupwelling Zn 6 £ 1025 d,e coastal areas and upwelling zones they are Al 1 £ 1024 d,e 250 g C m22 yr21 and 420 g C m22 yr21, respect- Cu a3 £ 1026 –0.006 d,e ively (Martin et al., 1987). However, because the Ni a2 £ 1025 –0.006 e a 27 open ocean constitutes 90% of the area of the Cd 5 £ 10 –0.005 d,e ocean, the bulk (80%) of the ocean’s annual a £ 26 Mn 4 10 –0.004 d,e carbon fixation occurs there rather than in coastal Ba a1 £ 1025 –0.01 d,e Mg a0.02 d and upwelling regions. Na a0.1 d Different types of phytoplankton dominate Sr a8 £ 1025 d primary production in the different marine Ti a1 £ 1025 d regimes. Diatoms perform roughly 75% of the Cr a2 £ 1026 d primary production that occurs in upwelling and coastal regions of the ocean but less than 35% of a Calculated from dry weight data using an average phytoplankton that taking place in the open ocean (Nelson et al., C content on a dry weight basis of 50%. b Redfield (1958). c Brzezinski (1985). d Martin and Knauer (1973). 1995). Phytoplankton biomass and primary pro- e Collier and Edmond (1984). f Sunda and Huntsman (1995a). duction in the open ocean are dominated instead 88 The Biological Pump

Figure 4 Distribution of primary production in the ocean (source ICMS, Rutgers University). by prokaryotic (Chisholm et al., Pacific, and the North Pacific subarctic) are never 1988; Liu et al., 1999; Steinberg et al., 2001). completely consumed in support of primary Outside of the tropics, levels of marine primary production, because low levels of iron limit productivity vary systematically throughout the phytoplankton growth (Martin and Fitzwater, year (Heinrich, 1962). Standing stocks of phyto- 1988; Martin et al., 1994; Coale et al., 1996; plankton and levels of primary production peak in Boyd et al., 2000). Diatoms, which, unlike other the spring following the onset of water column dominant members of the phytoplankton, require stratification and the increase in available light. silicon for growth, are often limited by low Depletion of nutrients in the stratified water concentrations of silicic acid in surface waters column in summer inhibits phytoplankton growth (Brzezinski and Nelson, 1996; Nelson and Dortch, and grazing by zooplankton reduces standing 1996). Growth of diazotrophic (N2 fixing) phyto- stocks. Some areas may experience a small bloom plankton such as the , Trichodes- of phytoplankton in the autumn when light levels mium, will be more susceptible to phosphorus and are still adequate and the onset of winter convec- iron limitation, of course, than to nitrogen tion and overturning injects nutrients into the limitation. Even the concentration of dissolved euphotic zone. CO2 in seawater (especially in the midst of a phytoplankton bloom) may limit instantaneous rates, although not ultimate levels, of primary production (Riebesell et al., 1993; Wolf-Gladrow 6.04.2.1.3 Nutrient limitation et al., 1999). The upper limit of primary production is set by the supply of nutrients (nitrogen, phosphorus, 6.04.2.2 Flocculation and Sinking silicon, iron) to the euphotic zone. Nitrogen inputs 6.04.2.2.1 Marine snow to the surface ocean may limit the primary productivity of the whole ocean over short time- The primary formation of biogenic particles in scales. Over timescales approaching and exceed- the euphotic zone represents the maximum ing the (1–5) £ 104 yr residence time of amount of material that may be transported phosphorus (Ruttenberg, 1993; Filippelli and into the deep ocean or sediments. In practice, Delaney, 1996), its inputs limit global ocean however, less than half of these particles survive primary productivity (Tyrrell, 1999). zooplankton grazing and microbial attack long Regionally and for different types of phyto- enough to be exported from the euphotic zone, plankton, the limitation of both the rate and and only a few percent endure to settle into the overall amount of primary production is more deep ocean and sediments (Martin et al., 1987). varied. Major nutrients in HNLC areas of the Material that reaches the deep ocean and ocean (such as the , the Equatorial seafloor does so not as individual phytoplankton Description of the Biological Pump 89 cells slowly meandering down towards the and TEP is required for the aggregation and sedi- bottom, but rather arrives as larger, rapidly mentation of diatoms out of the water column sinking particles (McCave, 1975; Suess, 1980; (Passow et al., 2001). Billett et al.,1983; Fowler and Knauer, 1986; Little is known about the chemical character- Alldredge and Gotschalk, 1989)thathave istics of the polymers responsible for particle traversed the distance between surface and aggregation in marine systems. They are com- deep in a matter of days (Billett et al., 1983; prised of acidic polysaccharides (Alldredge et al., Asper et al., 1992). 1993; Mopper et al.,1995) and proteins (Long and These larger particles, known collectively as Azam, 1996). The component of “marine snow” (Alldredge and Silver, 1988), are TEP contains glucose, mannose, arabinose, formed either by zooplankton, which produce xylose, galactose, rhamnose, glucuronate, and mucous feeding structures and fecal pellets, or by O-methylated sugars (Janse et al., 1996; Holloway the physical coagulation of smaller particles and Cowen, 1997), and is generally rich in (McCave, 1984; Alldredge et al., 1993). Coagu- deoxysugars (Mopper et al., 1995). Very little lation is the more important of the two formation else is known about the specific composition of pathways. The bulk of the organic material TEP, and virtually nothing is known of its reaching the deep sea does so as aggregated structural characteristics (Holloway and Cowen, phytoplankton that has not been ingested by 1997; Schumann and Rentsch, 1998; Engel and zooplankton (Billett et al., 1983; Turner, 2002). Passow, 2001). Sinking rates of marine snow are orders of Exopolymer particles are formed from dis- magnitude greater than those of unaggregated solved organic matter (DOM) and continue to phytoplankton cells (Smayda, 1970; Shanks and scavenge DOM as they grow, providing a Trent, 1980; Alldredge and Gotschalk, 1989). The mechanism for the biological pumping of DOM shorter transit time from surface to bottom for into the deep sea (Engel and Passow, 2001). TEP aggregated particles results in the enhanced transport of carbon, nitrogen, phosphorus, silicon, also contains carbon and nitrogen in proportions and other materials to the deep sea and sediments exceeding Redfield ratios (Mari et al., 2001; Engel despite the fact that marine snow particles are sites et al., 2002), providing a mechanism for pumping of elevated rates of decomposition and nutrient of carbon in excess of what would be predicted regeneration. Intense colonization and hydrolysis from the availability of nitrogen. of the particles by bacteria (Smith et al., 1992; Bidle and Azam, 1999, 2001) and breakup and consumption of the particles by zooplankton 6.04.2.2.3 Sinking (Steinberg, 1995; Dilling et al., 1998; Dilling and Alldredge, 2000) reduce the vertical flux of Sinking rates of solitary phytoplankton cells are materials to the seafloor. only about a meter per day (Smayda, 1970). Particles that sink this slowly require over a year to reach the benthos of the relatively shallow , and ten years to reach the 6.04.2.2.2 Aggregation and exopolymers abyssal ocean floor. Given the rapid rates of microbial decomposition of organic material in Coagulation requires the success of two activities: the collision of particles and their the ocean and the abundance of zooplankton subsequent joining to form an aggregate. In the grazers, it is virtually impossible for such a slowly ocean, particles collide due to processes such as sinking particle to reach the seafloor. Sinking rates of marine snow, however, are shear, Brownian motion, and differential settling 21 (Kepkay, 1994). The probability of particles greater than 100 m d (Shanks and Trent, 1980; attaching following a collision is controlled by Alldredge and Gotschalk, 1989). Transit time to the the physical and chemical properties of the deep in this case is days to weeks, which agrees with particles’ surfaces (Alldredge and Jackson, observations of a close temporal coupling between 1995). The probability of sticking is greatly surface production and seafloor sedimentation (e.g., enhanced by exopolymers produced by phyto- Billett et al.,1983; Asper et al.,1992). plankton and bacteria (Alldredge and McGillivary, It may easily be argued then that particle flux is 1991; Alldredge et al., 1993; Passow, 2000; controlled by rates of particle aggregation and Engel, 2000). sinking, perhaps more than it is controlled by These exopolymers, known as transparent exo- overall levels of primary production. For example, polymer particles (TEPs; Passow, 2002), turn out year-to-year variability in carbon export to deep to be important for the transport of material to the waters correlates more strongly with the size of deep. The formation of rapidly sinking aggregates the dominant primary producer than with year-to- is controlled more by TEP abundance than by year variations in levels of carbon fixation (Boyd phytoplankton concentrations (Logan et al., 1995) and Newton, 1995). 90 The Biological Pump 6.04.2.3 Particle Decomposition and portion of the sinking organic flux (Turner, 2002). Repackaging The second way they reduce the particle flux is by actively breaking up aggregates into smaller Organic matter in the ocean rapidly decomposes particles. At stations off Southern California, for and there is intense recycling of elements even instance, the average overnight increase in the within the euphotic zone. The flux of particulate number of aggregates per liter by 15% was organic carbon (POC) in the ocean decreases attributable to the fragmentation of larger particles exponentially with depth below the euphotic zone by swimming euphausiids (Dilling and Alldredge, (Figure 5; Martin et al., 1987). New production 2000). constitutes, on average, only 20% of the total The relative impact of zooplankton grazing on primary production in the sea (Harrison, 1990; primary production decreases with increasing Laws et al., 2000). Mediating this decomposition production levels; the proportion of primary and recycling are zooplankton and heterotrophic production that is consumed by zooplankton bacteria (Cho and Azam, 1988; Smith et al., 1992; decreases exponentially as productivity levels Steinberg, 1995; Dilling et al., 1998; Dilling and increase (Calbet, 2001). This supports the obser- Alldredge, 2000). Bacteria and zooplankton vation that the ratio of export production to total diminish the sinking particulate flux by both production is higher in areas of high productivity. consuming particles and converting them back to Globally, ,12% of marine primary production, or CO2 and dissolved materials, and by converting 5.5 Pg C (0.5 Pmol C), is consumed by mesozoo- large, sinking particles into smaller particles with plankton each year (Calbet, 2001). reduced or nonexistent sinking rates. Although the bulk of particles are broken down in the surface ocean, midwater processes 6.04.2.3.2 Bacterial hydrolysis are also important. Midwater decomposition of sinking particles deflates the regional vari- Bacterial hydrolysis plays a major role in the ability in fluxes of POM. The POC flux range, decomposition of sinking and suspended matter in 0.5–12 g C m22 yr21, among different regions in the ocean. Bacterial organic carbon has been the Atlantic at 125 m is compressed by 85% to observed to make up over 40% of the total POC in 0.5–2.4 g C m22 yr21 by a depth of 3,000 m due the water column, and the proportion of the to the biological consumption and repackaging of sinking flux of carbon utilized by bacteria may particles at depth (Anita et al., 2001). be equal to 40–80% of the surface primary production in near-shore areas (Cho and Azam, 1988). Marine aggregates have been shown to contain high concentrations of hydrolytic exo- 6.04.2.3.1 Zooplankton grazing enzymes, such as proteases, and polysaccharidases Zooplankton may reduce the sinking flux of like glucosidases (Smith et al., 1992). Turnover biogenic particles in the ocean in two ways. The times of organic components of marine aggregates first is by grazing upon particles which reduces the due to hydrolysis may be short, on the order of et al. total amount of particulate organic material fraction of a day to a few days (Smith , 1992). (POM) in the water column and shifts its Bacterial proteases have also been shown to occurrence from large, fast-sinking aggregates to enhance the dissolution rates of in smaller fecal pellets which constitute only a minor aggregates (Bidle and Azam, 2001). Much of the hydrolyzed material is not taken up by the bacteria attached to the particles but instead joins the pool of DOM present in the water (Smith et al., 1992).

6.04.2.3.3 Geochemistry of decomposition Ingestion of POM by zooplankton results in the respiration and excretion of a portion of that POM as CO2,NH4, dissolved organic nitrogen (DON), phosphate, and dissolved organic phosphorus (DOP). Assimilation efficiencies of organic matter for zooplankton grazing on phytoplankton range from 10% to 40% (Ryther, 1969; Michaels and Figure 5 Sinking fluxes of C in the open ocean. Curve Silver, 1988). Zooplankton do not assimilate shown is the exponential relationship (flux ¼ 1.53 significant quantities of silicon from diatoms (z/100)20.858) shown for carbon flux in the northeast consumed, leaving regeneration of silicic acid to Pacific (after Martin et al., 1987). be mediated strictly by opal dissolution rates. Description of the Biological Pump 91 6.04.2.4 Sedimentation and Burial and nitrogen for primary production (Clark et al., 1998; Zehr and Ward, 2002). DOM may assemble Of the 50–60 Pg C (4–5 Pmol C) fixed into into colloidal and particulate material that can sink organic material in the surface ocean each year as well as scavenge other material to form marine (Shuskina, 1985; Martin et al., 1987; Field et al., snow (Alldredge et al., 1993; Kepkay, 1994; Chin 1998) and the 16 Pg C (1.3 Pmol C) exported to et al., 1998). DOM is also a large reservoir of thedeepsea(Falkowski et al.,2000), only carbon in the ocean, containing at least an order of ,0.16 Pg C (0.13 Pmol C) reach the seafloor magnitude more carbon than the other organic (Hedges and Keil, 1995). Each year roughly carbon reservoirs in the ocean (Kepkay, 1994). 0.16 Pg C (13 Tmol C) are actually preserved in Although the origins of DOM have not been ocean sediments (Hedges and Keil, 1995). Thus, fully detailed, phytoplankton can serve as the the rate of accumulation of organic carbon by dominant source of DOM to the ocean. Actively ocean sediments is 0.5% of that of carbon fixation growing secrete DOM (Biddanda in the surface ocean. and Benner, 1997; Soendergaard et al., 2000; The accumulation of sedimentary organic Teira et al., 2001) and the polysaccharide com- carbon in the ocean varies greatly from place position of phytoplankton exudates resembles that to place, with sediments of the high molecular weight fraction of DOM accounting for the bulk of the organic carbon (Aluwihare and Repeta, 1999). Phytoplankton buildup (Hedges and Keil, 1995; de Haas et al., DOM is also released during grazing by zoo- 2002). About 94% of the sedimentary organic plankton (Strom et al., 1997). Organic matter, carbon that is preserved in the oceans is buried such as mucus, on phytoplankton cell surfaces on continental shelves and slopes (Hedges and may also be hydrolyzed by bacteria and released Keil, 1995). This leaves only 6% of the total as DOM (Smith et al., 1995). sedimentary organic carbon to accumulate in the The exact composition of marine DOM is open ocean. Since the open ocean, due to its vast unknown. It has, as of early 2000s, been shown area, plays host to 80% of the annual primary to contain carbohydrates, which consist largely production, the accumulation of only 6% of the of polysaccharides, and amino acids, amides (such organic carbon suggests an overall preservation as ), phosphorus esters, and phosphonates efficiency of 0.02% in the open ocean. On the et al. et al. continental shelves and slopes, this preservation (Benner , 1992; McCarthy , 1997; Clark efficiency is, by comparison, large at 1.4%. et al.,1998; Amon et al., 2001). Microbial Many reasons have been suggested for the degradation could play a role in setting the regional differences in the accumulation and composition of DOM in the ocean (Amon et al., preservation of organic carbon in the sediments. 2001). Differences in the flux of organic particles to the One feature that has great relevance to the seafloor due to differences in overhead primary importance of DOM to the is its production levels, rates of aggregation and sink- enrichment in carbon over the Redfield proportion. ing, and depth of the water column may contribute The C : N : P ratios of high molecular weight DOM to a higher preservation efficiency. The oxygen are on the order of 350 : 20 : 1 (Kolowith et al., content of bottom waters has also been suggested 2001). Carbon enrichment is also observed for as important, although a correlation between bulk DOM (Kaehler and Koeve, 2001). Part of this burial efficiency and bottom water oxygen con- enrichment in carbon may be due to the enhanced centration is not seen (Hedges and Keil, 1995) and remineralization of phosphorus and nitrogen from rates of organic matter hydrolysis by bacteria may DOM (Clark et al., 1998; Kolowith et al., 2001), as still be high, even under anoxic conditions (e.g., suggested by an increase in the C : P ratio of DOM Arnosti et al., 1994). The long-term preservation with depth (Kolowith et al., 2001). DOM may also of organic material in sediments may be tied to the just simply be produced with high C : N and C : P sorption of the organic to ratios. TEPs which form from DOM precursors surfaces (Hedges and Keil, 1995), although the (Alldredge et al., 1993; Chin et al., 1998) have of the associations and the rates at which high C : N ratios (Engel and Passow, 2001; they occur have not been closely detailed. Mari et al., 2001). The polysaccharides that eventually form TEP are exuded by phytoplankton and may represent excess photosynthate (Engel, 6.04.2.5 Dissolved Organic Matter 2002), carbohydrates, and lipids formed when nutrient limitation shuts off the supply of, for DOM has not been as intensively studied as example, the nitrogen needed for the synthesis of other aspects of the biological pump perhaps nitrogen-containing compounds such as proteins because DOM does not sink. However, DOM does (Morris, 1981). play an active role in the biological pump in at The importance of the carbon-enriched DOM least three ways. Much DOM can be utilized pool as a reservoir in the carbon cycle hinges upon biologically and may directly provide phosphorus both the turnover time and amount of the carbon 92 The Biological Pump therein. Estimates for the amount of dissolved dissolved CO2 increase most rapidly just organic carbon (DOC) in the ocean vary, although below the euphotic zone, associated with the an estimate places the size of the pool at 200 Pg C bulk of the decomposition of POC (Martin et al., (Kepkay, 1994), which is more comparable to the 1987; Anita et al., 2001). Deep-water concen- 750 Pg of carbon present in the atmosphere than it trations of dissolved CO2 are higher than is to the 3.6 £ 104 Pg C deep-sea reservoir of DIC surface-water concentrations, and older deep (Sundquist, 1993). The average age of marine waters contain more dissolved CO2 than younger DOM is ,6,000 yr (Williams and Druffel, 1987). ones (Broecker and Peng, 1982). The bulk (,70%) of the DOM is low molecular Much of the current scientific interest in the weight (Benner et al., 1992) and relatively resistant biological pump revolves around the impact it has to microbial degradation (Bauer et al., 1992; Amon on levels of CO2 in the atmosphere and, and Benner, 1994). High molecular weight com- subsequently, on climate. This biological fixation pounds that are quickly turned over by bacterial of carbon into organic matter through photosyn- decomposition (Amon and Benner, 1994) make up thesis lowers the concentration of dissolved CO2 the remaining 30% of the DOM pool. in surface waters and thus allows for the influx of CO2 from the atmosphere. This fixed carbon may then be exported to deeper waters or the sediments before it decomposes back to CO , maintaining 6.04.2.6 New, Export, and Regenerated 2 the observed gradient in dissolved CO2 concen- Production trations between waters of the surface and deep Not all of the primary production in the ocean (Figure 1). Atmospheric concentrations of CO2 feeds carbon into the biological pump. The vast are thus lower for the given size of the oceanic portion of carbon fixed globally each year in the DIC reservoir than they would be in the absence of euphotic zone is remineralized by zooplankton this biological transport of carbon to the deep. If and bacteria in the euphotic zone and converted all life in the ocean were to die off and the ocean and atmosphere came to equilibrium with respect straight back to CO2 and dissolved nutrients. These recycled nutrients may then be used to fuel to CO2, concentrations of CO2 in the atmosphere further carbon fixation. would rise by ,140 matm (Broecker, 1982), It has long been recognized (Dugdale and which is a remarkable 50% of the pre-Industrial Goering, 1967) that a portion of the primary interglacial value of 280 matm. production (regenerated production) is supported Details of the influence of the biological pump by nutrients regenerated in the euphotic zone, and on the distribution and cycling of carbon in the another portion (new production) is supported by ocean and the controlling factors are discussed nutrients imported into the euphotic zone through later in this chapter. We briefly discuss the upwelling, river inputs, nitrogen fixation, or relationship of the biological pump to the cycling atmospheric deposition. The ratio of new to total of other elements. This is by no means an primary production in the ocean, known as the exhaustive overview of the biological shuffling f-ratio (Eppley and Peterson, 1979), is generally of elements throughout the ocean, but instead a higher in upwelling environments than it is in highlight of several elements of particular bio- oligotrophic regions of the ocean (Harrison, 1990; geochemical interest. Laws et al., 2000). On average, ,20% of the total global marine primary production is new pro- 6.04.3.1.2 Nitrogen duction, although the range of values from region to region is ,0.07–0.7 (Laws et al.,2000). Of all the elements playing an important role in the regulation of the biological pump, nitrogen is the one with the most complex biologically mediated cycling. Nitrogenous species taking 6.04.3 IMPACT OF THE BIOLOGICAL PUMP part in productivity range from N , which may be ON GEOCHEMICAL CYCLING 2 fixed into a more universally biologically available 2 6.04.3.1 Macronutrients form by nitrogen fixing bacteria, to NO3 , which 2 þ 6.04.3.1.1 Carbon follows from the production of NO2 from NH4 through nitrification (Figure 6). The denitrification The influence of the biological pump on the pathway sequentially results in the transformation 2 distribution of DIC carbon in the ocean may of NO2 to N2O and N2. In addition to inorganic serve as a rough model for the influence it has nitrogen species are dissolved organic forms, such on the distribution of a score of other elements. as amides, urea, free amino acids, amines Low concentrations of DIC are observed in (McCarthy et al., 1997), which can be utilized surface waters (Figure 1) due to the uptake of biologically, although to varying degrees. 2 dissolved CO2 (and perhaps HCO3 ; Raven, Most of the primary production in the ocean is 1997) by phytoplankton. Concentrations of supported by dissolved inorganic nitrogen (DIN) Impact of the Biological Pump on Geochemical Cycling 93 diatoms were responsible for the bulk of the nitrogen fixation occurring in the ocean (Capone et al., 1997). However, direct measurements of the rates of nitrogen fixation by Trichodesmium, coupled with knowledge of their distribution and abundance, fell significantly short of nitrogen fixation rates calculated from geochemical budgets (Gruber and Sarmiento, 1997). It has now been discovered that free-living, unicellular cyanobac- teria are expressing the genes for the nitrogen- fixing enzyme, nitrogenase (Zehr et al., 2001), and may be contributing considerably to fixed nitrogen budgets in the ocean. It has been suggested that rates of nitrogen fixation in the modern ocean are limited by the availability of iron. The iron requirement of Trichodesmium, however, turns out to be much lower than previously estimated. Instead, it is the availability of phosphate that controls the upper limit of nitrogen fixation in the modern ocean (San˜udo-Wilhelmy et al., 2001). Figure 6 Microbial transformations of the nitrogen The demonstration of the phosphorus limitation cycle. Pathways depicted are: 1—N2 fixation; 2—DIN of nitrogen fixation by cyanobacteria supports the assimilation; 3— regeneration; 4—nitrifica- notion that over geologic time phosphorus ulti- tion; 5—nitrate/nitrite reduction; and 6—denitrification. mately limits productivity (Tyrrell, 1999). When cyanobacteria face a shortfall of nitrogen, they fix nitrogen to meet their demands. This influx of new that has recycled in the euphotic zone, as nitrogen to nitrogen-limited systems allows them suggested by the average marine f-ratio of 0.2 to draw down levels of phosphate until the system (Laws et al., 2000). This further suggests that the is phosphorus limited. Under phosphorus limi- bulk of the primary production in the ocean relies þ tation, nitrogen fixation is curtailed (San˜udo- on NH4 , since that is the predominant recycled Wilhelmy et al., 2001). Thus, while nitrogen form of nitrogen. Extrapolating from the Redfield may often be limiting to instantaneous rates of C : N ratio of 6.6 and the average f-ratio of 0.2 and carbon fixation, as is frequently the case in the the overall estimate of primary production in the modern ocean, over long time periods the input of ocean each year of 4–5 Pmol C (Shuskina, 1985; phosphorus to the oceans sets the upper attainable Martin et al., 1987; Field et al., 1998) yields limit for net primary production (Tyrrell, 1999). 0.7 Pmol of particulate organic nitrogen (PON) The control of nitrogen fixation by phosphorus produced each year, 0.1 Pmol of which is exported availability makes the stand out out of the euphotic zone. against all of the other global biogeochemical Nitrogen fixation. The two aspects of the cycles. In an astounding twist, the biological nitrogen cycle having the greatest impact on the demand for one element relative to the other biological pump are nitrogen fixation and deni- controls the input flux of one of the elements. trification. The first provides a mechanism for Confirmation of this comes in the form of Redfield drawing on the extensive atmospheric pool of N2 ratios, since the N : P ratio in both the seawater gas in support of primary production. The second reservoir and the marine organic matter output is provides a pathway for DIN to be converted back same (16 : 1) and significantly greater than unity. to N2 gas and removed from the ocean system. Due to mass balance requirements, assuming In the ocean ,28 Tg N (2 Tmol N) are fixed steady state over long timescales, the only way each year (Gruber and Sarmiento, 1997). Nitrogen this can be the case is if the ratio of N : P inputs to fixation accounts for about half of the new nitrogen the ocean is also 16 : 1. It is far more likely that the used in primary production (Karl et al., 1997). Only equality of N : P ratios of inputs, outputs, and prokaryotic can fix nitrogen, leaving reservoir is due to the grand control of nitrogen this, at least in the ocean, in the hands of the fixation by phosphate than it is likely to be due to cyanobacteria and out of the hands of eukaryotic remarkable coincidence. such as diatoms, dinoflagellates, and cocco- lithophorids (unless they are hosting diazotrophic 6.04.3.1.3 Phosphorus symbionts). Until recently it was believed that the filamentous, colony-forming cyanobacterium Tri- Although there is a tendency to consider chodesmium and cyanobacterial symbionts in dissolved inorganic phosphorus (DIP, measured 94 The Biological Pump as soluble reactive phosphate; Strickland and dissolved silicic acid to the euphotic zone are, by 32 Parsons, 1968) as being simply PO4 , DIP exists comparison, 120 Tmol Si, mostly from rivers as a considerable number of species. At seawater (5 Tmol Si) and upwelling (115 Tmol Si; Tre´guer 2 pH, DIP is predominantly H2PO4 (87%) and only et al., 1995). Half of the biogenic silica produced 32 2 12% PO4 and 1% H2PO4 (Greenwood and each year dissolves in the upper 100 m of the Earnshaw, 1984). There are also numerous water column (Nelson et al.,1995), and a further dissolved organic forms of phosphorus that are 47% dissolves in the deep ocean and seafloor, for taken up by phytoplankton and used to fuel a net deposition of 6 Tmol (3% of surface primary production in the ocean. production) each year (Tre´guer et al., 1995). One interesting aspect of the Opal accumulation on the seafloor.Opalpreser- is that, unlike the cases for the other major nutrient vation efficiencies are generally highest in pro- elements in the ocean, the phosphorus cycle faces ductive environments (e.g., 6% in the permanently the complexity of containing sinks that are not open ocean zone of the Southern Ocean versus 0.4% mediated by biological activities. For example, in the oligotrophic North Atlantic). Given numbers dissolved phosphate may scavenge onto iron or such as these, the traditional view of opal accumu- manganese oxyhydroxide particles associated lation in the sediments is that it is closely linked to with hydrothermal activity or react with opal production in overlying waters (e.g., Broecker during the circulation of water through mid-ocean and Peng, 1982), but many factors besides opal ridge hydrothermal systems (Fo¨llmi, 1996). The production govern opal preservation. Opal dissol- exact values are poorly known, but it is estimated ution rates more than double for every 10 8Crisein that the scavenging of phosphate by the hydro- temperature (Kamitani, 1982). Aggregation may thermal oxyhydroxides may constitute up to 50% reduce rates of silicon regeneration from diatoms of the removal flux of phosphorus from the ocean (Bidle and Azam, 2001). Additionally, the fraction (e.g., Froelich et al., 1982; Berner et al., 1993). of produced silica that reaches the correlates Another large inorganic sink for dissolved phos- with high seasonality and high ratios of carbon phorus is the precipitation of authigenic phos- export to production (Pondaven et al.,2000), which phate, which account for somewhere between also tend to be areas where organic matter is formed about 10% and 40% of the removal flux. Removal in blooms and is transported quickly to the of phosphorus as sedimenting POC by comparison sediments as large aggregates. is thought to be 20–50% of the output of The impact of the appearance of the diatoms on phosphorus from the ocean. the marine . The silica cycle is an Roughly 5 Tg P (0.2 Tmol P) are removed from excellent example of how much the biological the ocean each year as both organic and inorganic pump can impact concentration and distribution phases. This is in reasonable balance with the of elements in seawater. Presently, surface roughly 5 Tg of reactive phosphorus being brought concentrations of silicic acid are low: the overall in to the system each year naturally. However, average silicic acid content of ocean waters is these natural sources constitute only half of the only 70 mM(Tre´guer et al., 1995). Prior to the modern-day input of phosphorus to the ocean appearance of the diatoms in the Early Tertiary (Froelich et al.,1982), anthropogenic inputs (Tappan and Loeblich, 1973), oceanic concen- having doubled the annual phosphate flux. trations of silicic acid ranged ,1,000 mM Si. The A doubling of phosphorus inputs to the ocean and radiolarians controlling the oceanic could have a significant impact on the biological inventory of silicic acid at that time possessed pump, although it should be noted that the ocean is neither the numbers nor the need to draw down already nitrogen limited and further inputs of silicic acid concentrations. The increased output of phosphorus will only exacerbate this. Shifts in biogenic silica from the ocean associated with the N : P ratios of surface waters may alter the ascension of diatoms, with their high affinity for structure of the phytoplankton community. For silicic acid and high cellular requirements for example, Phaeocystis grows poorly under high- silicon, resulted in a precipitous drop in the silicic phosphate conditions and may see its numbers acid content of ocean waters over the Late declining. and Paleocene (Figure 7), stabilizing in the Eocene to the ,100 mM values that have 6.04.3.1.4 Silicon held ever since (Siever, 1991). Since sponges and radiolarians are not great Dissolved silicic acid is required for the growth players in particle flux, the rise of the diatoms must of diatoms, which deposit opal (amorphous, have profoundly altered the partitioning of silicic hydrated silica) in their cell wall and dominate acid between surface and deep. The familiar the production of opal in the modern-day ocean nutrient-type distribution may have only existed (Lisitzin, 1972). Silica (240 Tmol) is produced by for the last 50–100 million years. The approximate diatoms in the surface ocean each year (Nelson 14-fold drop in silicic acid concentration also et al., 1995; Tre´guer et al., 1995). Total inputs of suggests that the residence time of silicon in Impact of the Biological Pump on Geochemical Cycling 95

Figure 7 Estimated average marine concentrations of silicic acid over the Phanerozoic (after Siever, 1991).

Figure 8 Ratios of silicic acid to nitrate with depth in the Pacific. Profiles shown are from: (a) the NE Pacific WOCE station 66 (478 330 N, 1458 330 W); (b) HOT station Aloha (228 450 N, 1588 000 W); (c) SE Pacific WOCE station 288 (388 600 S, 888 000 W); and (d) SE Pacific WOCE station 241 (538 200 S, 768 360 W). the ocean dropped from 2 £ 105 yr to today’s nitrate close to the ,1 : 1 molar ratio of utilization 1.5 £ 104 yr. by diatoms, but Si : N ratios increase with depth Excessive pumping of silicon. Silicic acid, by below the euphotic zone (Figure 8). virtue of being regenerated from silica instead of Impact of iron on silicon pumping. Iron limitation from relatively labile POM, is regenerated more of diatoms increases the pumping of silicon (relative deeply than the other major nutrients (Figure 1; to nitrogen and carbon) to deeper waters. Iron- Dugdale et al., 1995). This decoupling between limited diatoms are inhibited in their utilization of silicic acid and the other nutrients, combined with nitrogen as a result of the iron requirement of nitrate the fact that not all phytoplankton utilize silicon, reductase (Geider and LaRoche, 1994). However, results in there being no Redfield relationship iron-limited diatoms continue to take up silicic acid, between silicon and carbon, nitrogen, or phos- although at lowered rates (De La Rocha et al.,2000). phorus. Upwelled waters contain silicic acid and As a result, the Si : N ratios of iron-limited diatoms 96 The Biological Pump maybeashighas2or3(Takeda, 1998; Hutchins and Bruland, 1998), much higher than the 0.8 of nutrient-replete diatoms (Brzezinski, 1985).

6.04.3.2 Trace Elements The biological pump influences, to varying degrees, the distribution of many elements in seawater besides carbon, nitrogen, phosphorus, and silicon. Barium, cadmium, germanium, zinc, nickel, iron, selenium, yttrium, and many of the REEs show depth distributions that very closely resemble profiles of the major nutrients. Addition- ally, beryllium, scandium, titanium, copper, zirconium, and radium have profiles where concentrations increase with depth, although the correspondence of these profiles with nutrient profiles is not as tight (Nozaki, 1997).

6.04.3.2.1 Barium Vertical profiles of dissolved barium (Ba2þ)in the ocean resemble profiles of silicic acid and (Figure 9; Lea and Boyle, 1989; Jeandel et al., 1996), suggesting that biological processes strongly influence barium distributions throughout the ocean. However, the strict incorporation of barium into biogenic materials is not the dominant means of Ba2þ removal from ocean waters. Despite the similarity between the profiles 2þ of Ba and Si(OH)4 which suggests a common removal phase, the amount of barium incorporated into diatom opal (,9 £ 1026 mol Ba per mol Si; Shemesh et al., 1988) cannot account for 24 2þ the 2 £ 10 mol Ba per mol Si(OH)4 slope 2þ (Jeandel et al., 1996) in the ocean. Ba appears 2þ instead to be mainly removed from seawater as Figure 9 Profiles of Ba , Si(OH)4, and alkalinity in the Indian Ocean (068 090 S, 508 550 E) (source Jeandel barite (BaSO4) formed in association with opal et al., 1996). and decaying organic material (Dehairs et al., 1980; Bishop, 1988). The exact mechanism for barite precipitation is unknown, but it is thought between the sedimentary fluxes of carbon and barite 22 (Dymond et al.,1992). Thus, barite accumulation that it forms in the SO4 -enriched microenviron- ments of decaying particles that may be, thus, rates have been used to infer past levels of export supersaturated with respect to barite (Dehairs production in the ocean (e.g., Dymond et al.,1992; et al., 1980). Paytan et al.,1996). Although the marine budget of barium is only approximately known, it does appear to be both 6.04.3.2.2 Zinc balanced and controlled by biogenic particle formation. Approximately 35 Gmol of Ba2þ are The profiles of dissolved zinc in the ocean are removed from surface waters every year (Dehairs also similar to those of Si(OH)4 (Figure 10; et al.,1980). Of this 35 Gmol, 60% settles as barite Bruland, 1980), but as with barium, the main and the rest is incorporated into or adsorbed onto removal phase for zinc is not opal. Less than 3% phases such as CaCO3 or SiO2 (Dehairs et al.,1980; of the zinc taken up by diatoms is deposited in Dymond et al.,1992). Barium (10–25 Gmol) is their opaline cell wall (Ellwood and Hunter, buried on the seafloor each year (Dehairs et al., 2000), and the Zn : Si ratio of acid-leached opal 1980). is much lower than that of dissolved Zn and Because barite forms in association with organic Si(OH)4 in the water column (Bruland, 1980; material, there is a tight correlation (R 6 ¼ 0.93) Collier and Edmond, 1984). Instead, most of Impact of the Biological Pump on Geochemical Cycling 97

2þ 2 2 Figure 10 Profiles of Zn, Cd , Ni, Cu, Si(OH)4,NO3 , and PO4 with depth in the North Pacific (after Bruland, 1980). 98 The Biological Pump the zinc removed from the surface ocean is bound Collier and Edmond, 1984). More labile com- up in POM. ponents of decaying particles have higher Cd : P Because zinc is required for the growth of ratios than bulk decaying particles (Knauer and phytoplankton, its availability affects the biologi- Martin, 1981). Box models of cadmium and cal pump. Although zinc limitation of an phosphorus cycling also require enhanced regen- entire phytoplankton community has never been eration of cadmium from particles for the replica- demonstrated, levels of dissolved zinc are often tion of observed cadmium distributions in the low enough to limit many taxa (Morel et al., 1994; ocean (Collier and Edmond, 1984). Sunda and Huntsman, 1995b; Timmermans et al., Cadmium is generally considered to be toxic to 2001). Zinc is an integral part of the enzyme, organisms and how the marine phytoplankton carbonic anhydrase (Morel et al., 1994), which utilize their cadmium is unknown. Cadmium may helps maintain an efficient supply of CO2 to substitute for zinc in carbonic anhydrase at times Rubisco. when zinc is limiting (Price and Morel, 1990; The impact of low concentrations of zinc on Lane and Morel, 2000). It is possible that phytoplankton growth varies. Some phytoplank- cadmium may play a role in polyphosphate ton substitute cobalt for zinc in many enzymes bodies, a form of cellular storage of phosphorus (Price and Morel, 1990; Sunda and Huntsman, that has been shown to contain significant 1995b; Timmermans et al., 2001)andcan quantities of elements such as calcium, zinc, and maintain maximal growth rates at low levels of magnesium (Ruiz et al., 2001). zinc. Calcification aids in the acquisition of DIC, The similarity between cadmium profiles and so phytoplankton, such as the cocco- nutrient profiles has been used as a means of lithophorids, are not as dependent on carbonic reconstructing past patterns of primary production anhydrase (and therefore zinc) to maintain high and nutrient cycling. Several attempts have been rates of carbon fixation (Sunda and Huntsman, made to reconstruct phosphate concentrations 1995b). Thus, while low levels of zinc may not from the Cd : Ca ratio of (Boyle, curtail overall levels of primary production, they 1988; Elderfield and Rickaby, 2000). Foramini- may shift the phytoplankton community structure fera substitute Cd2þ for Ca2þ in the lattice of their away from the diatoms and towards the cocco- tests, and the ratio of Cd : Ca incorporation lithophorids (Morel et al.,1994; Sunda and varies with the Cd : Ca ratio of seawater (Boyle, Huntsman, 1995b; Timmermans et al., 2001), 1988). The Cd : Ca ratio of foraminifera in the which will greatly impact the ratio of organic C to Southern Ocean suggests that phosphorus in CaCO3 of particles sinking to the deep sea. surface waters was not as heavily utilized by phytoplankton during the last glacial maximum (LGM), suggesting lower levels of primary production relative to nutrient flux into the 6.04.3.2.3 Cadmium euphotic zone at that time (Elderfield and Like barium and zinc, cadmium also shows a Rickaby, 2000). nutrient-like distribution in the ocean (Figure 10; Bruland, 1980; Lo¨scher et al., 1998), more closely 2 6.04.3.2.4 Iron mirroring those of the labile nutrients, NO3 and 2 PO4 , than those of Si(OH)4 and alkalinity. Of all the trace elements whose distributions are Cadmium is taken up by phytoplankton and affected by the biological pump, iron is the one that incorporated into organic material, accounting has the most profound impact on the workings of the 2 for the similarity of its profile to that of NO3 and biological pump. Iron plays a key role in many of the 2 PO4 . Cadmium may also be adsorbed onto the crucial enzymes in biological systems, such as surfaces of phytoplankton (Collier and Edmond, superoxide dismutase, ferredoxin, and nitrate 1984). reductase (Geider and LaRoche, 1994) that evolved Cadmium is taken up by phytoplankton slightly at a time when the oceans were low in oxygen and, 2 preferentially to PO4 (Lo¨scher et al., 1998; therefore, high in dissolved iron. As a result marine Elderfield and Rickaby, 2000). Waters with low phytoplankton have a heavier demand for iron 2 2 PO4 concentrations thus have lower Cd : PO4 relative to the present-day availability of dissolved 2 ratios (0.1 nmol Cd per mmol PO4 ) than waters iron in the ocean. Growth of phytoplankton and rates 2 2 with higher PO4 concentrations where Cd : PO4 of photosynthesis are frequently limited by the lack 2 may approach 0.4 nmol Cd per mmol PO4 of iron in surface waters (Martin and Fitzwater, (Elderfield and Rickaby, 2000). Additionally, 1989). 2 surface water Cd : PO4 ratios drop over the One of the major inputs of iron to the ocean development of the in the Southern comes from the dissolution of Fe(II) from wind- Ocean (Lo¨scher et al., 1998). borne continental dust deposited on the surface of Cadmium also regenerates preferentially from the ocean (Zhuang et al., 1990). In oxygenic decomposing particles (Knauer and Martin, 1981; environments, such as surface ocean waters, Fe(II) Quantifying the Biological Pump 99 will quickly be oxidized to the insoluble form Particle flux is frequently extrapolated from the Fe(III) and removed from seawater. Fe(II) is also measurement of new production in surface waters taken up by phytoplankton as well as complexed by (Dugdale and Goering, 1967). Export of POM out ligands exuded into the water by marine organisms of surface waters or into the deep sea may also be to prevent its precipitation as Fe(III). estimated directly through its collection in sedi- Profiles of dissolved iron in seawater show the ment traps (e.g., Martin et al., 1987; Anita et al., influence of both biotic and abiotic processes. At 2001). Export or sedimentation of POM may also stations in the Northeast Pacific, dissolved iron be estimated from disequilibria between two concentrations are low in surface waters, reflect- nuclides (e.g., 238U and 234Th, and 230Th and ing biological uptake. Iron concentrations also 231Pa) that are scavenged by particles of different show peak values at depth, corresponding to degrees (Buesseler et al., 1992; Kumar et al., the (Martin and Gordon, 1993; Franc¸ois et al.,1997). 1988), suggesting the abiotic reduction of Fe(III) Some methods for quantifying the biological back to the soluble Fe(II). pump focus the relationship between the biologi- Vast tracks of the ocean, such as the equatorial cal pump and CO2. POC flux measurements or Pacific, the Northeast Pacific subarctic, the estimates of nutrient removal from surface waters Southern Ocean, and even parts of the California may be used in conjunction with various ocean upwelling zone do not have sufficient supplies of models to estimate the impact of the biological iron to fully support phytoplankton growth (Martin pump on atmospheric concentrations of CO2 and Fitzwater, 1988; Martin et al., 1994; Coale (Sarmiento and Toggweiler, 1984; Sarmiento et al., 1996; Hutchins and Bruland, 1998; Boyd and Orr, 1991). Others have focused on the et al., 2000). In these areas, macronutrients such as importance of the ratio of POC to CaCO3 to the nitrogen and phosphorus are rarely depleted. sequestering of CO2 in the ocean (Anita et al., Attention has turned to these HNLC areas as sites 2001; Buitenhuis et al., 2001). where further primary production (and the associ- ated drawdown of atmospheric CO2) could occur. Increased supplies of dust stimulating the biologi- 6.04.4.1 Measurement of New Production cal pump in HNLC regions may be responsible for The method in most widespread use for low atmospheric CO2 concentrations during gla- quantifying the biological pump hinges upon the cial times (Martin and Gordon, 1988; Watson et al., ideas that the surface ocean is at steady state on 2000). Artificially stimulating the biological pump annual timescales with respect to the nitrogen by seeding HNLC areas with chelated iron has been budget and that nitrogen predominantly limits proposed as a means of pumping the 90 matm of phytoplankton growth in the ocean. In such a CO2 in the atmosphere (put there by humans) into system, the amount of productivity that is the deep sea, although there is not much consensus exported from the euphotic zone must be equal as to the effectiveness of such an endeavor to the amount of productivity that is fuelled by the (Chisholm et al., 2001). input of allocthonous or “new” nitrogen to the euphotic zone (Dugdale and Goering, 1967). According to this definition, “new production” is 6.04.4 QUANTIFYING THE BIOLOGICAL one that is supported from dissolved nitrogen PUMP upwelled into the euphotic zone or fixed from N2 into PON, and export production is taken to be There are many different ways to quantify the equal to new production. For the sake of biological pump. Total levels of carbon fixation in measurement, the above definition of new pro- surface waters may be estimated in bottle duction has been simplified even further. Experi- 14 incubations from the uptake of CO2 by phyto- mentally, new production is taken to be equal to 2 2 plankton (Steemann Nielsen, 1952) or from the the production that uses NO3 (and NO2 also but deviation of oxygen isotopes from their terrestrial more in theory than in practice) as its nitrogen þ mass fractionation line (Luz and Barkan, 2000). source, as opposed to NH4 or any of the organic At the other end of the biological pump, forms of nitrogen. sedimentary accumulation rates of organic carbon It should be pointed out that measurement of may be measured. In between, the impact of the new production is the measurement of the feeding and vertical migration activities of mid- maximum amount of productivity that can be water organisms on the flux of particles may be exported without running the system down with investigated (e.g., Steinberg, 1995; Dilling et al., respect to the annual supply of nutrients to the 1998). However, because of the current interest in euphotic zone. It may not always be appropriate to the CO2 pumping capacity of the biological pump, assume that new production and export production we concentrate here on the methods used to are equal; the fluxes will be equal only in systems estimate export production and particle flux from that are not evolving. The validity of the the euphotic zone. assumption that new production equals export 100 The Biological Pump production depends on to what degree and over to depth via the biological pump. Measurement of what timescale the nitrogen cycle in surface new production divulges no information concerning waters is at steady state. the depth of decomposition of the POM formed or At the moment the nitrogen cycle in the ocean is the ratio of POC to CaCO3 oftheexportedparticles not at steady state on the decadal timescale. In (Anita et al.,2001). For instance, CO2 from material the last few decades, the N : P ratio of many oceanic decomposed beneath the euphotic zone but above waters has changed (Pahlow and Riebesell, 2000; the maximum depth of winter mixing will be Emerson et al., 2001) and near-shore waters have ventilated straight back out the atmosphere. shifted towards silicon limitation from nitrogen CaCO3 formation, a feature that is not common to limitation (Conley et al., 1993) due to anthropo- all phytoplankton, diminishes the efficiency of CO2 genic inputs of DIN to these systems and an drawdown with primary production (see below; increase in the extent of nitrogen fixation. On very Buitenhuis et al. (2001)). long timescales, the nitrogen cycle is not at steady state either, but responds to changes in the oceanic inventory of phosphorus, which ultimately governs the rate of nitrogen fixation (Tyrrell, 1999). These 6.04.4.2 Measurement of Particle Flux imbalances may not be large enough on a yearly timescale to affect the estimate of export Means more direct than the measurement of production from new production, but no new production exist for the estimation of particle assessment of this has been made. fluxes into the deep. Moored or free-floating traps Another assumption open to question is that the may be used to collect sinking particles (e.g., 2 Martin et al., 1987). Alternatively, particle flux rate of NO3 uptake adequately represents the rate of uptake of new forms of nitrogen. This may be estimated from particle-reactive nuclides (e.g., Buesseler et al., 1992). Particle flux may simplification has come about for two reasons. 2 also be estimated from the consumption of oxygen The first is that the uptake rate of NO by 3 (associated with the decomposition of sinking phytoplankton can be measured with reasonable 2 POM) in waters below the surface layer (e.g., ease by tracking the uptake of 15N-labeled NO 3 Jenkins, 1982). (Dugdale and Goering, 1967). The second is that 2 NO3 is not produced in the euphotic zone to any significant degree and so its presence there can only be as a result of upwelling or atmospheric 6.04.4.2.1 Sediment traps deposition. The form of DIN released during the death, The collection of particles in sediment traps, þ decay, and grazing of phytoplankton is NH4 , while perhaps the most direct way of measuring which is also the most easily utilizable form of DIN the sinking flux of POM, is a method not free from to phytoplankton. Oxidized forms of DIN, such as a certain amount of controversy. Sediment traps 2 22 þ NO3 and NO2 , must be reduced to NH4 by both over-collect and under-collect particles. nitrifying bacteria such as Nitrosomonas, Nitro- Comparison of 234Th accumulating in a suite of bacter, Nitrospira, and Proteobacteria (Zehr and sediment traps with 234Th fluxes expected from Ward, 2002) prior to assimilation by phytoplank- U–Th disequilibria in the upper 300 m of the ton. The classical view is that nitrification does not ocean suggested that the particle collection occur in the euphotic zone due to the inhibition of efficiency of these traps ranged from 10% to nitrifying bacteria by light (Zehr and Ward, 2002). 1,000% (Buesseler, 1991). Traps deployed in the Eukaryotic phytoplankton are also thought to deep ocean also show a considerable variability in outcompete nitrifying bacteria for the supplies of trapping efficiencies (e.g., Scholten et al., 2001). þ 2 NH4 in surface waters. Thus, NO3 found in the Despite the magnitude of these biases, there is no euphotic zone must have had its origins outside of generally applied method for correcting fluxes the euphotic zone, in deeper waters, in rivers or using particle-reactive isotopes (Anita et al., agricultural runoff, or from atmospheric deposition 2001). and is taken as the sole representative of new Zooplankton actively swimming into sediment nitrogen in the euphotic zone (Dugdale and traps also serve as a source of error in flux Goering, 1967). measurements. It is impossible to differentiate 2 Of course, NO3 is not likely to be the only form these “swimmers” from zooplankton that have of allocthonous nitrogen in the euphotic zone. settled passively into the cup as part of the sinking þ Concentrations of NH4 in upwelled water are not POM flux. Swimmers may constitute as much as zero, although they are much lower than those of a quarter of the POC collected by the trap 2 NO3 . DON may also serve as a significant source (Steinberg et al., 1998) and are generally removed of new nitrogen to the euphotic zone. from trap material prior to analysis. This intro- There is one last word of caution concerning the duces minimal error into the trap estimates of POC 2 use of NO3 uptake to estimate the transport of CO2 flux, as detrital zooplankton likely only comprise Quantifying the Biological Pump 101 ,2% of the total organic matter sinking into the reactive as 230Th, is highest in areas of high traps (Steinberg et al., 1998). particle flux (Anderson et al., 1990). Thus, sediments in high flux areas exhibit 231Pa/230Th ratios in excess of the initial production ratio of 0.093, and sediments accumulating slowly exhibit 6.04.4.2.2 Particle-reactive nuclides ratios less than 0.093 (Anderson et al., 1990). 231 230 Radionuclides in the uranium decay series serve Pa/ Th ratios have been used to infer changes as useful tracers of particle flux. One type of these in productivity and sediment accumulation rates tracers consists of a soluble parent nuclide and a between the present-day interglacial and the 4 particle-reactive daughter. These soluble nuclide– LGM, ,2 £ 10 yr ago (Kumar et al., 1993; particle-reactive pairs include 238U–234Th, Franc¸ois et al., 1997). 234U–230Th, and 235U–231Pa. The half-life of the parent exceeds the mixing time of the ocean and its distribution throughout the ocean is uniform. 6.04.4.2.3 Oxygen utilization rates Once the soluble parent isotope decays to the particle-reactive daughter, the daughter is sca- The distribution of oxygen in ocean waters venged onto particulate material. contains information about primary production. In systems with no particle scavenging, the For example, the amount of excess oxygen activities of the parent and daughter nuclide will present in the seasonal thermocline in the Pacific be in secular equilibrium. What is seen instead is was long ago used to suggest that 14C-based that the activities of 234Th, 230Th, and 231Pa are estimates of primary production were severely lower in surface waters than those of their parents underestimating levels of primary production in (Figure 11). The difference in the activities of the ocean (Shulenberger and Reid, 1981). Oxygen parent and daughter is a measure of the uptake of deficiencies in deeper waters have been used to daughter onto particles (Buesseler et al., 1992). estimate levels of export production (Jenkins, With the help of a model of particle scavenging, 1982). fluxes of the particle-reactive daughter may be The supplies of oxygen to waters below the estimated from its vertical distribution (e.g., Coale euphotic zone are primarily physical: advection and Bruland, 1985; Buesseler et al., 1992). If the and mixing. Removal of oxygen from these waters ratio of the particle-reactive nuclide to POC or takes place through the oxidation of organic PON is known, then the calculated flux of nuclide matter (Equations (2) and (3)): can be converted to an estimate of particle flux (Buesseler et al., 1992). O2 þ CH2O ! CO2 þ H2O ð2Þ Relative estimates of particle flux may also be made from the ratio of two particle-reactive 2 2 2O2 þ NH3 þ OH ! NO3 þ 2H2O ð3Þ nuclides, such as 230Th and 231Pa, which are scavenged onto particles to different degrees. By measuring rates of ventilation and the degree The half-lives of these isotopes are much of oxygen undersaturation in deeper waters, an larger than their residences times in the ocean estimate of the rates of oxygen utilization (OUR) (,104 yr versus tens to hundreds of years), and may be made and integrated to yield the total thus there is no significant radioactive decay that amount of oxygen consumed beneath the euphotic occurs in the water column. The extent of the zone each year (Jenkins, 1982). From this number scavenging of 231Pa, which is not as particle a flux of POC and PON may be calculated. Estimates of export production based on OURs are reasonably in line with the amount of new production that could be supported from measured 2 fluxes of NO3 into surface waters (Jenkins, 1988). One advantage of using the OUR method for estimating export fluxes is that it integrates over larger spatial and temporal scales than do estimates based on sediment traps and nuclide fluxes. Also unlike the estimates of new pro- 2 duction based on NO3 uptake, OURs are directly coupled to the recycling of CO2 via particle decomposition and thus a more direct measure of the impact of the biological pump on atmospheric CO2. In practice, however, the measurement of 2 NO3 uptake is less technically challenging and is Figure 11 Profiles of 234Th (circles) and 238U carried out much more frequently than are (diamonds) in the upper ocean (after Buesseler, 1991). estimates from nuclides and OURs. 102 The Biological Pump 6.04.5 THE EFFICIENCY OF THE control both the proportion and total amount of BIOLOGICAL PUMP organic carbon pumped to the deep sea.

6.04.5.1 Altering the Efficiency of the Biological Pump 6.04.5.1.1 In HNLC areas There is much talk concerning the “efficiency” of the biological pump. Is it pumping as much The biological pump in HNLC areas is not carbon to the deep sea as it could be? The general operating at full efficiency on at least two counts. consensus is that it is not operating at its full Phytoplankton growth is curtailed by the lack of capacity, and this is generally meant to imply that availability of trace metals such as iron, and so globally, the nitrate flux into the euphotic zone is concentrations of the major nutrients, nitrogen and not fully consumed in support of marine primary phosphorus, are not drawn down and carbon production. For example, Broecker (1982) has fixation does not occur to its maximum possible suggested that if all of the nutrients supplied to extent (Martin and Fitzwater, 1988). In addition, ocean surface waters were consumed by phyto- iron limitation heavily impacts the larger phyto- plankton, such as diatoms, that are important to plankton, the atmospheric CO2 content would drop by ,130 ppmv. particle flux. When a phytoplankton community is released from iron limitation, diatom growth is A simplistic estimate of how much more CO2 the biological pump could draw out of the stimulated more strongly than the growth of the atmosphere gives a false impression of our other phytoplankton (Cavender-Bares et al., 1999; understanding of the system. One hint that there Lam et al., 2001). Given all of this, the addition of iron to HNLC is a considerable decoupling of primary pro- waters should result in higher levels of carbon duction and nutrient drawdown from levels of fixation, increased growth of diatoms, local export production is the regional variability in the drawdown of CO , and enhanced export of carbon ratio of export to total production. Levels of 2 to the deep sea. Iron-addition experiments in export production are higher in more highly bottles and in situ on the mesoscale unequivocally productive areas than in low-productivity areas, support the first two points. Addition of iron to as would be expected. However, a greater HNLC waters results in madly blooming phyto- proportion of the total primary production is plankton (Martin and Fitzwater, 1989; Martin exported from mesotrophic and eutrophic regions et al., 1994; Coale et al., 1996; Boyd et al., 2000; than from oligotrophic areas of the ocean. The Strass, 2002). Chlorophyll concentrations may ratio of export production to total production may quadruple (e.g., Strass, 2002) and carbon fixation be as low as 0.12 in oligotrophic areas but as high rates may triple (e.g., Coale et al., 1996) over the as 0.4 in highly productive regimes (Anita et al., first few days after iron addition (Figure 12). 2001). Alternatively, support for the drawdown of There is a lesson to be learnt here with regard to CO2 following iron addition has been mixed. In spurring the biological pump into action. Most of the IronEx-I, the first mesoscale iron experiment, iron talk regarding increasing the efficiency of the addition did not result in a marked drop of CO2 con- biological pump focuses only on productivity and centrations in surface waters of the equatorial nutrient drawdown, but there is much more to the Pacific (Watson et al., 1994). Some CO2 was drawn biological pump than nutrient drawdown. The main down during IronEx-II, also in the equatorial question asked is: Can we stimulate an increase in Pacific, but the amount removed from surface primaryproduction(andbyextensionexport waters was still not enough to prevent these production) by adding iron to the ocean? From the recently upwelled, high CO2 waters from out- point of view of ascertaining whether or not gassing CO2 to the atmosphere (Coale et al., 1996). increased glacial dust could have stimulated the The Southern Ocean, however, is a better ecosys- biological pump and could have been responsible tem than the equatorial Pacific for stimulating CO2 for the lower levels of CO2 then present in the drawdown with iron. Both large-scale iron- atmosphere, this is an appropriate question. But addition experiments that have been carried out from the point of view of effectively sequestering there (SOIREE and EisenEx-1) produced a signifi- anthropogenic carbon in the deep sea via the cant lowering (25–30 matm) of the CO2 content of biological pump, this is a very narrow approach. surface waters (Boyd et al., 2000; Watson et al., Carbon sequestration via the biological pump really 2000; Strass, 2002). cannot be considered as a practical activity until we The one key critical point that all of the iron- have a better predictive understanding of not only enrichment experiments have failed to show is an primary production but also particle aggregation increase in the export of carbon to the deep sea. Even and sinking, zooplankton grazing and microbial the Southern Ocean experiments, where a CO2 hydrolysis, and organic carbon to CaCO3 pro- drawdown occurred, did not result in an increase in duction and rain rates. These are the factors that export flux. Sediment traps set out at a depth of The Efficiency of the Biological Pump 103

Figure 12 2 P et al. Responses to Fe addition to HNLC waters: chlorophyll, carbon fixation, NO3 , and CO2 (after Boyd , 2000; Watson et al., 2000).

100 m (below the fertilized patch) in the SOIREE iron-addition experiments. The first is that there is experiment collected a similar amount of POC to no increase in export flux, but instead an increase in traps stationed outside of the patch (Boyd et al., the flow of carbon through the 2000). In the EisenEx-1 experiment, there was no (microzooplankton and bacteria) which would increase in sinking POC over the three-week result in the regeneration of CO2 and nutrients in window of the experiment (Strass, 2002). the euphotic zone. Microzooplankton biomass and There are many possible reasons for the lack of an bacterial activities were seen to increase in observed increase in export flux in the mesoscale many of the mesoscale iron-addition experiments 104 The Biological Pump (e.g., Boyd et al.,2000; Strass, 2002), suggesting and lowers the alkalinity (approximately equal to 22 2 that some fraction of the carbon fixed through iron 2[CO3 ] þ [HCO3 ]) of the surface waters (Zeebe addition was being shunted into the microbial loop and Wolf-Gladrow, 2001). Loss of alkalinity instead of being pumped into deeper waters. decreases the capacity of surface waters to play Alternatively, the sediment traps used may simply host to dissolved CO2. In the surface ocean, under have not been deployed for long enough or in the an atmosphere with a CO2 of right location to catch the carbon sinking out as a 350 matm (19 matm less than the 2001 values; result of the iron-induced bloom. Keeling and Whorf, 2001), the precipitation of 1 mol of CaCO3 results in the release of 0.6 mol of CO2 concentrations (Ware et al.,1992; Frankignoulle et al., 1994). 6.04.5.1.2 Through changes in community The coccolithophorid, , con- composition tains, on average, 0.433 mol of CaCO3 (present as Changes in community composition should scales covering the cell) for every mole of organic profoundly affect the efficiency of the biological carbon (Buitenhuis et al., 2001). Thus, for every pump. Shifting productivity from small cells, mole of CO2 fixed into organic matter such as photosynthetic picoplankton, which do by coccolithophorids such as E. huxleyi, roughly not efficiently sink, to large cells, such as diatoms, 0.26 mol of CO2 will be released due to the capable of aggregating into particles that sink at formation of CaCO3. A shift, therefore, from hundreds of meters a day will result in the export production being carried out by cocco- pumping of more carbon to the deep. However, lithophorids to a noncalcareous phytoplankton there is finite amount of productivity that may be such as diatoms will result in an increased exported without running the nutrients in the drawdown of carbon for the same amount of system down to zero. In a steady-state system, primary production. The fact that carbon fixation shifting the community structure cannot increase by phytoplankton that calcify is only 75% as the total amount of export production above the effective at removing CO2 as carbon fixation done level of new production. However, a shift in by phytoplankton that do not precipitate CaCO3 community structure may export carbon deeper complicates estimates of CO2 drawdown from the into the water column before it decomposes, surface ocean POC export flux (Anita et al., 2001). thus delaying the return of that carbon to the surface ocean by a significant number of years. A shift in community composition may also be 6.04.5.1.3 By varying the C : N : P ratios of important to the biological pump if it is from a sinking material calcareous to a noncalcareous phytoplankton, as The C : N : P ratios of sinking POM are not precipitation of CaCO3 diminishes theP ocean’s fixed; phosphorus and nitrogen are preferentially ability to hold dissolved CO2. DIC ( CO2)is regenerated from sinking POM, which allows for present in seawater as several species, dissolved the sequestering of more carbon in the deep ocean CO2, , and the dissociated forms, per unit nitrogen or phosphorus than if Redfield , and carbonate ion: ratios remain unaltered during the decomposition X p 2 22 of organic matter. It was remineralization ratios CO2 ¼ H2CO3 þ HCO3 þ CO3 ð4Þ calculated from DIC and nutrient concentrations along that initially suggested that the * (where H2CO3 represents the sum of dissolved C : N of sinking particles might be higher than the CO2 and carbonic acid). The ratio of the three value of 6.6 proposed by Redfield (Takahashi species varies largely with pH, with the undisso- et al., 1985). material also showed ciated forms dominating only at low pH and that the C : N and C : P ratios of sinking particles 2 HCO3 favored at the typical seawater pH of increase with depth, from near Redfield values of 8.3 (Zeebe and Wolf-Gladrow, 2001). The 6.6 and 106 at the base of the euphotic zone to * saturation concentration of H2CO3 is controlled ,11 and 180 by 5,000 m (Martin et al., 1987). 2 22 in large part by the amount of HCO3 and CO3 Since the above initial observations, evidence 2 present in . Removal of HCO3 or has mounted for the preferential remineralization 22 CO3 during CaCO3 formation results in a of nitrogen and phosphorus out of POM relative to lowering of the saturation concentration of carbon. Ocean models with continuous vertical dissolved CO2, and therefore the outgassing of resolution support the preferential release of CO2 from solution to the atmosphere. nitrogen and phosphorus from sinking POM in Precipitation of CaCO3 both produces CO2 line with the estimates of Martin et al. (Shaffer, (Equation (5)): 1996). Bacteria have been shown to more rapidly et al. 2þ 2 degrade PON than POC (Verity , 2000). Ca þ 2HCO3 ! CaCO3 þ H2O þ CO2 ð5Þ Increases in the C : N and C : P of sinking particles The Biological Pump in the Immediate Future 105 have also been observed at station ALOHA near sinking of intact phytoplankton from the euphotic Hawaii (Christian et al., 1997). Differences in the zone. In lower latitude, oligotrophic-area standing C : N ratio of suspended POC (6.4) and less stocks of phytoplankton and zooplankton are not as buoyant POC trapped at the pycnocline (8.6) in decoupled in time and there is no clear window the Kattegat, just east of the North Sea, further within which a high density of phytoplankton cells suggest that nitrogen is remineralized more may aggregate, escape predation, and sink to the rapidly than carbon at significant levels even in deep. the euphotic zone (Olesen and Lundsgaard, 1995).

6.04.6 THE BIOLOGICAL PUMP IN THE 6.04.5.1.4 By enhancing particle transport IMMEDIATE FUTURE While sediment trap evidence supports the idea What may we expect out of the biological pump that the flux of organic carbon (JC org) to depth (z) in the future? We cannot currently predict how can be estimated from levels of primary pro- ocean biology will respond to climate warming duction (PP) (Equation (6)): (Sarmiento et al., 1998), but there are other questions that we may begin to ask. Will, for : 1:77 20:68 JCorg¼ 0 1PP z ð6Þ instance, the biological pump respond to the rise in surface ocean concentrations of DIC tied to the the correlation between POC flux and overlying rise in atmospheric CO2 fuelled by anthropogenic levels of primary production (R 2 ¼ 0.53; Anita emissions? What impact will inputs of agricultural et al., 2001) is not strong. The ratio of POC flux (at fertilizers have on the biological pump? What role 125 m) to primary production also shows great will artificial stimulation of the biological pump variability from region to region, ranging from through deliberate ocean fertilization play in the 0.08 to 0.38 in the Atlantic Ocean alone (Anita sequestration of excess CO2? et al., 2001). There are many reasons that an increased downward transport of POC out of the euphotic zone does not directly follow from an increase in 6.04.6.1 Response to Increased CO2 productivity. Export fluxes are also controlled by Will the biological pump respond to the the packaging of smaller more numerous particles increase CO2? Increasing concentrations of CO2 into larger less numerous aggregates and by the do not appear to have the significant impact on the rate at which the resulting aggregates sink. The C : N or C : P ratios of phytoplankton (Burkhardt depth to which POC exported from the euphotic et al., 1999) that is necessary if CO2 is to stimulate zone is transported must be determined by the productivity in the ocean. However, possibly the balance between the carbon content of the particle, rate of carbon-rich exudation by phytoplankton the rate of microbial hydrolysis, and the sinking and the production of TEP will rise with rate of the particle. Carbon fixed by marine increasing CO concentrations (Engel, 2002). It diatoms, which are large and adept at aggregating, 2 may also be possible for CO2 levels to increase the stands a better chance at being exported to the proportion of carbon diverted into the biological deep than carbon fixed by very small, unsinking pump at the expense of grazing and microbial food picoplankton. webs. Riebesell et al. (1993) suggested that carbon The mode of productivity is also crucial to the fixation by phytoplankton may be rate limited by sinking flux of POC. Systems exporting POC in the diffusive flux of CO2 to the cell which delivers pulses (e.g., following the periodic formation of 2 2 CO2 too slowly relative to NO3 and PO4 to take phytoplankton blooms in temperate and high up carbon, nitrogen, and phosphorus in Redfield latitude areas) export a much higher proportion proportions. of their total production than do systems, such as The diffusive flux of CO2 to the surface of a the oligotrophic central gyres, where carbon fixation F phytoplankton cell ( CO2) is controlled not only by is a more static, steady process (Lampitt and Anita, the diffusivity of CO (D) and the radius of the cell 1997). This results partly from the influence of 2 (r), but the concentration gradient of CO2 between particle concentration on aggregation and partly the aqueous medium and the interior of the cell from zooplankton population dynamics. In temper- (Ce –Ci) and the rate constant for the conversion of ate areas, zooplankton reproduction cannot begin HCO2 to CO (k0) as the carbonate system until after the onset of the spring phytoplankton 3 2 equilibrates following CO2 removal (Equation bloom, leading to a lag in the increase in (7)): zooplankton population (and subsequently in the grazing pressure exerted by zooplankton) behind ffiffiffiffiffir ffi that of the increase in phytoplankton numbers FCO ¼ 24prD 1 þ p ðCe 2 CiÞð7Þ (Heinrich, 1962) allowing for the aggregation and 2 D=k0 106 The Biological Pump

If the diffusive flux of CO2 is controlling the the removal of silicon from these systems and a delivery of CO2 to the carbon-fixing enzyme, shift in the phytoplankton community structure Rubisco, then an increase in the dissolved CO2 away from diatoms. Given that rivers are signifi- 2 2 content of seawater should stimulate phytoplank- cant sources of new NO3 ,PO4 , and Si(OH)4 to ton growth rates. In particular, an increase in the global ocean, this effect is expected to spread dissolved CO2 concentrations should stimulate further throughout the sea. growth rates in the larger (i.e., aggregate forming) A shift in community structure associated with size classes and thus should also result in a silicic acid depletion may sharply reduce the shunting of carbon out of the mouths of zoo- amount of carbon delivered to deep waters and plankton and into the biological pump (Riebesell sediments via the biological pump. Diatoms, et al., 1993). which are the only major phytoplankton group It is not clear that the acquisition of carbon is requiring silicic acid, are relatively large and the rate-limiting step for photosynthesis even in notable for aggregating and sinking. Diatoms feed the larger phytoplankton cells. Phytoplankton a greater portion of the organic matter they are not limited to the passive diffusion of CO2 produce into the biological pump than do most to the cell surface for carbon acquisition. other classes of phytoplankton. Currently, diatoms 2 HCO3 may be taken up directly by phyto- perform more than 75% of the primary production plankton and used as a source of CO2 for that occurs in high nutrient and coastal regions of photosynthesis (Korb et al., 1997; Nimer et al., the ocean (Nelson et al., 1995), the exact areas that 2 1997; Raven, 1997; Tortell et al., 1997). The will be impacted by this input of additional NO3 2 2 considerably greater abundance of HCO3 than and PO4 and the exact areas where carbon tends dissolved CO2 in seawater would imply that to make it down into the sediments. carbon-fixation rates of marine phytoplankton One possible further impact of the decline of the are not CO2 limited, and increasing concen- diatoms lies in the identity of their probably trations of CO2 will not have an impact on successor. If a CaCO3-producing phytoplankter, rates of primary production. such as coccolithophorids, steps in to utilize the 2 2 NO3 and PO4 the silicon-starved diatoms leave behind, the CO2 pumping efficiency of the 6.04.6.2 Response to Agricultural Runoff biological pump will decline (Robertson et al., 1994) even if the same amount of organic carbon Inputs of agricultural fertilizers are having continues to be removed to the deep sea and more than one impact on the biological pump. 2 2 sediments each year. A shift in the NO3 :PO4 : Si(OH)4 of natural waters is causing a shift in the phytoplankton community structure that should impact the biological cycling of carbon in aquatic systems 6.04.6.2.2 Shifts in export production and (Conley et al., 1993). Additionally, a recent deep ocean C : N : P shift in the C : N : P ratios of deeper waters and There is evidence for anthropogenic pertur- an increase in export production have been bations increasing the biological pumping of observed for the northern hemisphere oceans carbon into deep waters. In both the North Pacific (Pahlow and Riebesell, 2000). and North Atlantic, for example, deep-water N : P ratios have increased since the early 1950s (Pahlow and Riebesell, 2000) possibly due to 6.04.6.2.1 Shift towards silicon limitation atmospheric deposition of anthropogenic nitrogen Since the 1850s, changes in land use patterns, and a subsequent shift towards phosphorus human population density, and extent of the use limitation in these areas. Concomitant with the of fertilizers have resulted in an increased flux of rise in N : P is an increase in the apparent oxygen 2 2 NO3 and PO4 to rivers, lakes, and the coastal utilization (AOU) in both the North Atlantic ocean (Conley et al.,1993) and even to the open and North Pacific, which suggests that levels of ocean through atmospheric deposition, for export production have also increased. This is example, of anthropogenic nitrous (Pahlow estimated to have resulted in the increased oceanic 21 and Riebesell, 2000). At the same time, large sequestration of 0.2 Pg C yr (Pahlow and freshwater systems, such as the North American Riebesell, 2000). Great Lakes, and coastal areas, such as the Adriatic and Baltic and the Mississippi River plume, have been shifting away from nitro- 6.04.6.3 Carbon Sequestration via Ocean gen or phosphorus limitation and towards silicon Fertilization and the Biological Pump limitation (Conley et al.,1993; Nelson and Dortch, 1996). The extra productivity fuelled on There have been many calls to sequester anthro- the extra nitrogen and phosphorus has resulted in pogenic CO2 in the deep ocean by stimulating References 107 primary production through the addition of REFERENCES literally tons of iron to HNLC surface waters. Alldredge A. L. and Gotschalk C. C. (1989) Direct obser- Patents have been taken out on the idea (e.g., vations of the mass flocculation of diatom blooms: Markels, 2001) and, counter to the recommen- characteristics, settling velocities and formation of diatom dations of the American Society of Limnology aggregates. Deep-Sea Res. 36, 159–171. and Oceanography (http://www.aslo.org/policy/ Alldredge A. L. and Jackson G. A. (1995) Aggregation in docs/oceanfertsummary.pdf), several companies marine systems. Deep-Sea Res. II 42, 1–7. Alldredge A. L. and McGillivary P. (1991) The attachment have been established to dispense carbon credits probabilities of marine snow and their implications for to industries willing to pay for ocean fertilization particle coagulation in the ocean. Deep-Sea Res. 38, (Chisholm et al., 2001). However, it remains 431–443. unclear whether ocean fertilization will ever Alldredge A. L. and Silver M. W. (1988) Characteristics, successfully transport carbon in the deep sea, dynamics, and significance of marine snow. Prog. Ocea- how long the transported carbon might remain out nogr. 20, 41–82. Alldredge A. L., Passow U., and Logan B. E. (1993) The of contact with the atmosphere, and what side- abundance and significance of a class of large, transparent effect large-scale fertilization will have on organic particles in the ocean. Deep-Sea Res. 40, marine geochemistry and ecology (Fuhrman and 1131–1140. Capone, 1991; Peng and Broecker, 1991a,b; Aluwihare L. I. and Repeta D. J. (1999) A comparison of the Sarmiento and Orr, 1991; Chisholm et al., 2001; chemical characteristics of oceanic DOM and extracellular DOM produced by marine algae. Mar. Ecol. Prog. Ser. 186, Lenes et al., 2001). 105–117. A perusal of the literature suggests that carbon Amon R. M. W. and Benner R. (1994) Rapid cycling of high- sequestration by is not a panacea molecular-weight dissolved organic matter in the ocean. for the anthropogenic carbon emissions that Nature 369, 549–552. increased atmospheric CO from 280 matm at Amon R. M. W., Fitznar H.-P., and Benner R. (2001) Linkages 2 among the bioreactivity, chemical composition, and diage- the start of the industrial revolution to 369 matm netic state of marine dissolved organic matter. Limnol. by December 2000 (Keeling and Whorf, 2001). Oceanogr. 46, 287–297. Ocean models suggest that enhanced ocean uptake Anderson R. F., Lao Y., Broecker W. S., Trumbore S. E., of carbon with iron fertilization of the Antarctic Hofman H. J., and Wolfli W. (1990) Boundary scavenging in 10 231 Ocean will at best draw down atmospheric CO by the Pacific Ocean: a comparison of Be and Pa. Earth 2 Planet. Sci. Lett. 96, 287–304. 70 matm if carried out continuously for a century Anita A. N., et al. (2001) Basin-wide particulate carbon flux in (Peng and Broecker, 1991a,b) and damp the the Atlantic Ocean: regional export patterns and potential for annual anthropogenic CO2 input to the atmos- atmospheric CO2 sequestration. Global Biogeochem. Cycles phere by less than 30% of current annual emission 15, 845–862. levels (Joos et al., 1991). Arnosti C., Repeta D. J., and Blough N. V. (1994) Rapid bacterial degradation of polysaccharides in anoxic marine Large-scale iron fertilization will have side- systems. Geochim. Cosmochim. Acta 58, 2639–2652. effects. If iron fertilization resulted in a complete Asper V. L., Deuser W. G., Knauer G. A., and Lohrenz S. E. drawdown of the nutrients available in the (1992) Rapid coupling of sinking particles fluxes Southern Ocean, for example, the average O2 between surface and deep ocean waters. Nature 357, content of deep waters will drop by 4–12% 670–672. (Sarmiento and Orr, 1991), with areas of anoxia Azam F. and Volcani B. E. (1981) Germanium–silicon interactions in biological systems. In Silicon and Siliceous cropping up in the Antarctic (Peng and Broecker, Structures in Biological Systems (eds. T. L. Simpson and 1991b) and Indian Oceans (Sarmiento and Orr, B. E. Volcani). Springer, pp. 69–93. 1991). Even small-scale patches of anoxia would Bauer J. E., Williams P. M., and Druffel E. R. M. (1992) 14C have a profound negative impact on the survival activity of fractions in the north- and distribution of metazoan fauna in the ocean central Pacific and Sargasso Sea. Nature 357, 667–670. Benner R., Pakulski J. D., McCarthy M., Hedges J. I., and and alter the balance of microbial transformations Hatcher P. G. (1992) Bulk chemical characteristics of of nitrogen between reduced and oxidized phases dissolved organic matter in the ocean. Science 255, (Fuhrman and Capone, 1991). 1561–1564. Berner R. A., Ruttenberg K. C., Ingall E. D., and Rao J. L. (1993) The nature of phosphorus burial in modern marine sediments. In Interactions of C, N, P and S Biochemical Cycles and ACKNOWLEDGMENTS Global Change (eds. R. Wollast, F. T. Mackenzie, and L. Chou). Springer, pp. 365–378. The author thanks U. Passow for a helpful Biddanda B. and Benner R. (1997) Carbon, nitrogen, and carbohydrate fluxes during the production of particulate and review, A. Engel, U. Passow, and A. Wischmeyer dissolved organic matter by marine phytoplankton. Limnol. for insightful conversation, A. Alldredge, Oceanogr. 42, 506–518. R. Collier, and L. Dilling for promptly sending Bidle K. D. and Azam F. (1999) Accelerated dissolution of reprints, A. Knoll and A. Yool for pointing out the diatom silica by natural marine bacterial assemblages. long-term silica cycle, W. Berelson, H. Elderfield Nature 397, 508–512. Bidle K. D. and Azam F. (2001) Bacterial control of silicon and N. McCave for helpful tips, and R. Banger regeneration from diatom : significance of bacterial and S. Bishop for finding everything the author ectohydrolases and specie identity. Limnol. Oceanogr. 46, could not. 1606–1623. 108 The Biological Pump

Billett D. S. M., Lampitt R. S., Rice A. L., and Mantoura R. F. Collier R. and Edmond J. (1984) The trace element C. (1983) Seasonal sedimentation of phytoplankton to the geochemistry of marine biogenic particulate matter. Prog. deep sea benthos. Nature 302, 520–522. Oceanogr. 13, 113–199. Bishop J. K. B. (1988) The barite-opal-organic carbon associ- Conley D. J., Schelske C. L., and Stoermer E. F. (1993) ation in oceanic particulate matter. Nature 332, 341–343. Modification of the of silica with Boyd P. and Newton P. (1995) Evidence of the potential eutrophication. Mar. Ecol. Prog. Ser. 101, 179–192. influence of planktonic community structure on the inter- de Haas H., van Weering T. C. E., and de Stigter H. (2002) annual variability of particulate organic carbon flux. Deep- Organic carbon in shelf seas: sinks or sources, processes and Sea Res. I 42, 619–639. products. Cont. Shelf Res. 22, 691–717. Boyd P. W., et al. (2000) A mesoscale phytoplankton bloom in Dehairs F., Chesselet R., and Jedwab J. (1980) the polar Southern Ocean stimulated by iron fertilization. Discrete suspended particles of barite and the barium Nature 407, 695–702. cycle in the open ocean. Earth Planet. Sci. Lett. 49, Boyle E. A. (1988) Cadmium: chemical tracer of deepwater 528–550. paleoceanography. Paleoceanography 3, 471–489. De La Rocha C. L., Hutchins D. A., Brzezinski M. A., and Broecker W. S. (1982) during glacial time. Zhang Y. (2000) Effects of iron and zinc deficiency on Geochim. Cosmochim. Acta 46, 1689–1705. elemental composition and silica production by diatoms. Broecker W. S. and Peng T.-H. (1982) Tracers in the Sea. Mar. Ecol. Prog. Ser. 195, 71–79. Eldigio Press. de Villiers S. (1999) Seawater strontium and Sr/Ca variability Bruland K. W. (1980) Oceanographic distributions of cad- in the Atlantic and Pacific oceans. Earth Planet Sci. Lett. mium, zinc, nickel, and copper in the North Pacific. Earth 171, 623–634. Planet. Sci. Lett. 47, 176–198. Dilling L. and Alldredge A. L. (2000) Fragmentation of marine Brzezinski M. A. (1985) The Si : C : N ratio of marine diatoms: snow by swimming macrozooplankton: a new process interspecific variability and the effect of some environmental impacting carbon cycling in the sea. Deep-Sea Res. I 47, variables. J. Phycol. 21, 347–357. 1227–1245. Brzezinski M. A. and Nelson D. M. (1996) Chronic substrate Dilling L., Wilson J., Steinberg D., and Alldredge A. (1998) limitation of silicic acid uptake in the western Sargasso Sea. Feeding by the euphausiid Euphausia pacifica and the Deep-Sea Res. II 43, 437–453. Calanus pacificus on marine snow. Mar. Ecol. Buesseler K. O. (1991) Do upper-ocean sediment traps provide Prog. Ser. 170, 189–201. and accurate record of particle flux? Nature 353, 420–423. Dugdale R. C. and Goering J. J. (1967) Uptake of new and Buesseler K. O., Bacon M. P., Cochran J. K., and Livingston H. D. regenerated forms of nitrogen in primary productivity. (1992) Carbon and nitrogen export during the JGOFS North Limnol. Oceanogr. 12, 196–206. Atlantic Bloom Experiment estimated from 234Th : 238Udis- Dugdale R. C., Wilkerson F. P., and Minas H. J. (1995) The equilibria. Deep-Sea Res. 39, 1115–1137. role of a pump in driving new production. Deep-Sea Buitenhuis E. T., van der Wal P., and de Baar H. J. W. (2001) Res. I 42, 697–719. Blooms of Emiliania huxleyi are sinks of atmospheric carbon Dymond J., Suess E., and Lyle M. (1992) Barium in deep-sea dioxide: a field and mesocosm study derived simulation. sediment: a geochemical for paleoproductivity. Global Biogeochem. Cycles 15, 577–587. Paleoceanography 7, 163–181. Burkhardt S., Zondervan I., and Riebesell U. (1999) Effect of Elderfield H. and Rickaby R. E. M. (2000) Oceanic Cd/P CO2 concentration on C : N : P ratio in marine phytoplank- ratio and nutrient utilization in the glacial Southern Ocean. ton: a species comparison. Nature 405, 305–310. Calbet A. (2001) Mesozooplankton grazing effect on primary Elderfield H. E. and Schultz A. (1996) Mid-ocean ridge production: a global comparative analysis in marine hydrothermal fluxes and the chemical composition of the . Limnol. Oceanogr. 46, 1824–1830. ocean. Ann. Rev. Earth Planet. Sci. 24, 191–224. Capone D. G., Zehr J. P., Paerl H. W., Bergman B., and Ellwood M. J. and Hunter K. A. (2000) The incorporation of Carpenter E. J. (1997) Trichodesmium: a globally significant zinc and iron into the of the marine diatom marine cyanobacterium. Science 276, 1221–1229. . Limnol. Oceanogr. 45, Cavender-Bares K. K., Mann E. L., Chisholm S. W., Ondrusek 1517–1524. M. E., and Bidigare R. R. (1999) Differential response of Emerson S., Mecking S., and Abell J. (2001) The biological equatorial Pacific phytoplankton to iron fertilization. Limnol. pump in the subtropical North Pacific Ocean: nutrient Oceanogr. 44, 237–246. sources, Redfield ratios, and recent changes. Global Chin W. C., Orellana M. V., and Verdugo P. (1998) Biogeochem. Cycles 15, 535–554. Spontaneous assembly of marine dissolved organic matter Engel A. (2000) The role of transparent exopolymer particles into polymer gels. Nature 391, 568–572. (TEP) in the increase in apparent particle stickiness (alpha) Chisholm S. W., Falkowski P. G., and Cullen J. J. (2001) Dis- during the decline of a diatom bloom. J. Plankton Res. 22, crediting ocean fertilization. Science 294, 309–310. 485–497. Chisholm S. W., Olson R. J., Zettler E. R., Goericke R., Engel A. (2002) Direct relationship between CO2 uptake and Waterbury J. B., and Welschmeyer N. A. (1988) A novel transparent exopolymer particle production in natural free-living prokaryote abundant in the oceanic euphotic phytoplankton. J. Plankton Res. 24, 49–53. zone. Nature 334, 340–343. Engel A. and Passow U. (2001) Carbon and nitrogen content of Cho B. C. and Azam F. (1988) Major role of bacterial in transparent exopolymer particles (TEP) in relation to their biogeochemical fluxes in the ocean’s interior. Nature 332, Alcian Blue absorption. Mar. Ecol. Prog. Ser. 219,1–10. 441–443. Engel A., Goldthwait S., Passow U., and Alldredge A. (2002) Christian J. R., Lewis M. R., and Karl D. M. (1997) Vertical Temporal decoupling of carbon and nitrogen dynamics fluxes of carbon, nitrogen, and phosphorus in the North in a mesocosm diatom bloom. Limnol. Oceanogr. 47, Pacific Subtropical Gyre near Hawaii. J. Geophys. Res. 102, 753–761. 15667–15677. Eppley R. W. and Peterson B. J. (1979) Particulate organic Clark L. L., Ingall E. D., and Benner R. (1998) Marine matter lux and planktonic new production in the deep ocean. phosphorus is selectively remineralized. Nature 393, 428. Nature 282, 677–680. Coale K. H. and Bruland K. W. (1985) 234Th:238U disequilibria Fagerbakke K. M., Norland S., and Heldal M. (1999) The within the California current. Limnol. Oceanogr. 30, 22–33. inorganic ion content of native aquatic bacteria. Can. Coale K. H., et al. (1996) A massive phytoplankton bloom J. Microbiol. 45, 304–311. induced by an -scale iron fertilization experiment Falkowski P., et al. (2000) The global carbon cycle: a of in the equatorial Pacific Ocean. Nature 383, 495–501. our knowledge of Earth as a system. Science 290, 291–296. References 109

Field C. B., Behrenfeld M. J., Randerson J. T., and Falkowski Kaehler P. and Koeve W. (2001) Marine dissolved organic P. (1998) Primary production of the biosphere: integrating matter: can its C : N ratio explain carbon overconsumption? terrestrial and oceanic components. Science 281, 237–240. Deep-Sea Res. I 48, 49–62. Filippelli G. M. and Delaney M. L. (1996) Phosphorus Kamitani A. (1982) Dissolution rates of silica from diatoms geochemistry of equatorial Pacific sediments. Geochim. decomposing at various temperatures. Mar. Biol. 68, 91–96. Cosmochim. Acta 60, 1479–1495. Karl D. M., Letelier R., Tupas L., Dore J., Christian J., and Fo¨llmi K. B. (1996) The phosphorus cycle, phosphogenesis Hebel D. (1997) The role of nitrogen fixation in biogeo- and marine phosphate-rich deposits. Earth Sci. Rev. 40, chemical cycling in the subtropical North Pacific Ocean. 55–124. Nature 388, 533–538. Fowler S. W. and Knauer G. A. (1986) Role of large particles in Keeling C. D. and Whorf T. P. (2001) Atmospheric CO2 the transport of elements and organic compounds through records from sites in the SiO air sampling network. the oceanic water column. Prog. Oceanogr. 16, 147–194. In Trends: A Compendium of Data on Global Change. Franc¸ois R., Altabet M. A., Yu E.-F., Sigman D. M., Bacon US Department of Energy, Oak Ridge National Laboratory, M. P., Frank M., Bohrmann G., Bareille G., and Labeyrie Information Analysis Center, Oak Ridge, L. D. (1997) Contribution of Southern Ocean surface-water TN. stratification to low atmospheric CO2 concentrations during Kepkay P. E. (1994) Particle aggregation and the biological the last glacial period. Nature 389, 929–935. reactivity of colloids. Mar. Ecol. Prog. Ser. 109, 293–304. Frankignoulle M., Canon C., and Gattuso J.-P. (1994) Marine Knauer G. A. and Martin J. H. (1981) Phosphorus and cadmium calcification as a source of carbon dioxide: positive feedback cycling in northeast Pacific waters. J. Mar. Res. 39, 65–76. of increasing atmospheric CO2. Limnol. Oceanogr. 39, Kolowith L. C., Ingall E. D., and Benner R. (2001) 458–462. Composition and cycling of marine organic phosphorus. Froelich P. N., Bender M. L., Luedtke N. A., Heath G. R., and Limnol. Oceanogr. 46, 309–321. De Vries T. (1982) The marine phosphorus cycle. Am. J. Sci. Korb R. A., Saville P. J., Johnston A. M., and Raven J. A. 282, 474–511. (1997) Sources of inorganic carbon for photosynthesis by Fuhrman J. A. and Capone D. G. (1991) Possible biogeochem- three species of marine diatom. J. Phycol. 33, 433–440. ical consequences of ocean fertilization. Limnol. Oceanogr. Kumar N., Gwiazda R., Anderson R. F., and Froelich P. N. 36, 1951–1959. (1993) 231Pa/230Th ratios in sediments as a proxy for past Geider R. J. and LaRoche J. L. (1994) The role of iron in changes in Southern Ocean productivity. Nature 362, phytoplankton photosynthesis, and the portential for iron 45–48. limitation of primary productivity in the sea. Photosyn. Res. Lam P. J., Tortell P. D., and Morel F. M. M. (2001) Differential 39, 275–301. effects of iron additions on organic and inorganic carbon Greenwood N. N. and Earnshaw A. (1984) Chemistry of the production by phytoplankton. Limnol. Oceanogr. 46, Elements. Pergamon. 1199–1202. Gruber N. and Sarmiento J. L. (1997) Global patterns of marine Lampitt R. and Anita A. N. (1997) Particle flux in deep seas: nitrogen fixation and denitrification. Global Biogeochem. regional characteristics and temporal variability. Deep-Sea Cycles 11, 235–266. Res. I 44, 1377–1403. Harrison P. J., Conway H. L., Holmes R. W., and Davis C. O. Lane T. W. and Morel F. M. M. (2000) A biological function (1977) Marine diatoms grown in chemostats under silicate or for cadmium in marine diatoms. Proc. Natl Acad. Sci. USA ammonium limitation: III. Cellular chemical composition 97, 4627–4631. and morphology of debilis, Skeletonema Laws E. A., Falkowski P. G., Smith W. O., Jr., Ducklow H., costatum, and Thalassiosira gravida. Mar. Biol. 43, 19–31. and McCarthy J. J. (2000) Temperature effects on export Harrison W. G. (1990) Nitrogen utilization in chlorophyll and production in the open ocean. Global Biogeochem. Cycles primary productivity maximum layers: an analysis based on 14, 1231–1246. the f-ratio. Mar. Ecol. Prog. Ser. 60, 85–90. Lea D. W. and Boyle E. (1989) Barium content of benthic Hedges J. I. and Keil R. G. (1995) Sedimentary organic matter foraminifera controlled by bottom water composition. preservation: an assessment and speculative synthesis. Mar. Nature 338, 751–753. Chem. 49, 81–115. Lenes J. M., et al. (2001) Iron fertilization and the Heinrich A. K. (1962) The life histories of planktonic animals Trichodesmium response on the West Florida shelf. Limnol. and seasonal cycles of plankton communities in the ocean. Oceanogr. 46, 1261–1277. J. Cons. Int. Explor. Mer. 27, 15–24. Levitus S., Burgett R., and Boyer T. (1994) World Ocean Atlas Holloway C. F. and Cowen J. P. (1997) Development of a 1994 Volume 3: Nutrients. US Department of Commerce. scanning confocal laser microscopic technique to examine Lisitzin A. P. (1972) Sedimentation in the World Ocean. Soc. the structure and composition of marine snow. Limnol. Econ. Paleontol. Mineral. Spec. Publ. 17, 218pp. Oceanogr. 42, 1340–1352. Liu H., Landry M. R., Vaulot D., and Campbell L. (1999) Hutchins D. A. and Bruland K. W. (1998) Iron-limited diatom growth rates in the central equatorial growth and Si : N uptake ratios in a coastal upwelling Pacific: an application of the fmax approach. J. Geophys. regime. Nature 393, 561–564. Res. 104, 3391–3399. Janse I., vanRijssel M., Gottschal J. C., Lancelot C., and Logan B. E., Passow U., Alldredge A. L., Grossart H. P., Gieskes W. W. C. (1996) Carbonhydrates in the North Sea and Simon M. (1995) Rapid formation and sedimentation during spring blooms of Phaeocystis: a specific fingerprint. of large aggregates is predictable from coagulation rates Aquat. Microb. Ecol. 10, 97–103. (half-lives) of transparent exopolymer particles (TEP). Jeandel C., Dupre´ B., Lebaron G., Monnin C., and Minster J.-F. Deep-Sea Res. II 42, 203–214. (1996) Longitudinal distributions of dissolved barium, silica Long R. A. and Azam F. (1996) Abundant -containing and alkalinity in the western and southern Indian Ocean. particles in the sea. Aquat. Microb. Ecol. 10, 213–221. Deep-Sea Res. I 43, 1–31. Lo¨scher B. M., de Jong J. T. M., and de Baar H. J. W. (1998) Jenkins W. J. (1982) Oxygen utilization rates in North Atlantic The distribution and preferential biological uptake of subtropical gyre and primary production in oligotrophic cadmium at 68W in the Southern Ocean. Mar. Chem. 62, systems. Nature 300, 246–248. 259–286. Jenkins W. J. (1988) Nitrate flux into the euphotic zone near Luz B. and Barkan E. (2000) Assessment of oceanic Bermuda. Nature 331, 521–523. productivity with the triple-isotope composition of dissolved Joos F. L., Sarmiento J. L., and Sigenthaler U. (1991) Estimates oxygen. Science 288, 2028–2031. of the effect of Southern Ocean iron fertilization on Mari X., Beauvais S., Lemee R., and Pedrotti M. L. (2001) atmospheric CO2 concentrations. Nature 349, 772–775. Non-reedfield C : N ratio of transparent exopolymeric 110 The Biological Pump

particles in the northwestern Mediterranean Sea. Limnol. exopolymer particles (TEP) and their role in the sedimen- Oceanogr. 46, 1831–1836. tation of particulate matter. Cont. Shelf Res. 21, 327–346. Markels M., Jr. (2001) Method of sequestering carbon dixoide Paytan A., Kastner M., and Chavez F. P. (1996) Glacial to with spiral fertilization. US Patent No. 6,200,530. interglacial fluctuations in productivity in the Equatorial Martin J. H. (1990) Glacial-interglacial CO2 change: the iron Pacific as indicated by marine barite. Science 274, hypothesis. Paleoceanography 5, 1–13. 1355–1357. Martin J. H. and Fitzwater S. E. (1988) Iron deficiency limits Peng T.-H. and Broecker W. S. (1991a) Dynamical limitations phytoplankton growth in the north-east Pacific subarctic. on the Antarctic iron fertilization strategy. Nature 349, Nature 331, 341–343. 227–229. Martin J. H. and Gordon R. M. (1988) Northeast Pacific iron Peng T.-H. and Broecker W. S. (1991b) Factors limiting the distributions in relation to phytoplankton productivity. reduction of atmospheric CO2 by iron fertilization. Limnol. Deep-Sea Res. 35, 177–196. Oceanogr. 36, 1919–1927. Martin J. H. and Knauer G. A. (1973) The elemental Pondaven P., Ragueneau O., Tre´guer P., Hauvespre A., composition of plankton. Geochim. Cosmochim. Acta 37, Dezileau L., and Reyss J. L. (2000) Resolving the “opal 1639–1653. paradox” in the Southern Ocean. Nature 405, 168–172. Martin J. H., Knauer G. A., Karl D. M., and Broenkow W. W. Price N. M. and Morel F. M. M. (1990) Cadmium and cobalt (1987) VERTEX: carbon cycling in the northeast Pacific. substitution for zinc in a marine diatom. Nature 344, Deep-Sea Res. 34, 267–285. 658–660. Martin J. H., et al. (1994) Testing the iron hypothesis in Quigley M. S., Santschi P. H., Hung C.-C., Guo L., and Honeyman B. D. (2002) Importance of acid polysaccharides ecosystems of the equatorial Pacific Ocean. Nature 371, 234 123–129. for Th complexation to marine organic matter. Limnol. McCarthy M., Pratum T., Hedges J., and Benner R. (1997) Oceanogr. 47, 367–377. Chemical composition of dissolved organic nitrogen in Raven J. A. (1997) Inorganic carbon acquisition by marine the ocean. Nature 390, 150–154. . Adv. Bot. Res. 27, 85–209. McCave I. N. (1975) Vertical flux of particles in the ocean. Redfield A. C. (1934) On the proportions of organic derivatives Deep-Sea Res. 22, 491–502. in sea water and their relation to the composition of McCave I. N. (1984) Size spectra and aggregation of suspended plankton. In James Johnstone Memorial Volume (ed. particles in the deep ocean. Deep-Sea Res. 31, 329–352. R. J. Daniel). University Press of Liverpool, pp. 177–192. Michaels A. F. and Silver M. W. (1988) Primary production, Redfield A. C. (1958) The biological control of chemical sinking fluxes and the . Deep-Sea Res. factors in the environment. Am. Sci. 46, 205–221. 35, 473–490. Riebesell U., Wolf-Gladrow D. A., and Smetacek V. (1993) Mopper K., Zhou J. A., Ramana K. S., Passow U., Dam Carbon dioxide limitation of marine phytoplankton growth rates. Nature 361, 249–251. H. G., and Drapeau D. T. (1995) The role of surface- Robertson J. E., Robinson C., Turner D. M., Holligan P., active carbohydrates in the flocculation of a diatom Watson A. J., Boyd P., Fernandez E., and Finsh M. (1994) bloom in a mesocosm. Deep-Sea Res. II 42, 47–73. The impact of a bloom on oceanic carbon Morel F. M. M., Reinfelder J. R., Roberts S. B., Chamberlain uptake in the northeast Atlantic during summer 1991. Deep- C. P., Lee J. G., and Yee D. (1994) Zinc and carbon co- Sea Res. I 41, 297–314. limitation of marine phytoplankton. Nature 369, 740–742. Ruiz F. A., Marchesini N., Seufferheld M., and Docampo R. Morris I. (1981) Photosynthetic products, physiological state, (2001) The polyphosphate bodies of Chlamydomonas and phytoplankton growth. Can. J. . Aquat. Sci. 210, reinhardtii possess a proton-pumping pyrophosphatase and 83–102. are similar to acidocalcisomes. J. Biol. Chem. 276, Nelson D. M. and Dortch Q. (1996) Silicic acid depletion and 46196–46203. silicon limitation in the plume of the Mississippi River: Ruttenberg K. C. (1993) Reassessment of the oceanic residence evidence from kinetic studies in spring and summer. Mar. time of phosphorus. Chem. Geol. 107, 405–409. Ecol. Prog. Ser. 136, 163–178. Ryther J. H. (1969) Photosynthesis and fish production in the Nelson D. M., Tre´guer P., Brzezinski M. A., Leynaert A., and sea. Science 166, 72–76. Que´guiner B. (1995) Production and dissolution of biogenic San˜udo-Wilhelmy S. A., Kustka A. B., Gobler C. J., Hutchins silica in the ocean: revised global estimates, comparison D. A., Yang M., Lwiza K., Burns J., Capone D. G., Raven with regional data and relationship to biogenic sedimen- J. A., and Carpenter E. J. (2001) Phosphorus limitation of tation. Global Biogeochem. Cycles 9, 359–372. nitrogen fixation by Trichodesmium in the central Atlantic Nimer N. A., Iglesias-Rodriguez M. D., and Merrett M. J. Ocean. Nature 411, 66–69. (1997) Bicarbonate utilization by marine phytoplankton Sarmiento J. L. and Orr J. C. (1991) Three-dimensional species. J. Phycol. 33, 625–631. simulations of the impact of Southern Ocean nutrient Nozaki T. (1997) A fresh look at element distribution in the depletion on atmospheric CO2 and ocean chemistry. Limnol. North Pacific. EOS Trans., AGU Electron. Suppl http:// Oceanogr. 36, 1928–1950. www.agu.org/eos_elec/97025e.html. Sarmiento J. L. and Toggweiler J. R. (1984) A new model for Olesen M. and Lundsgaard C. (1995) Seasonal sedimentation P the role of the oceans in determining atmospheric CO2. of autochthonous material from the euphotic zone of a Nature 308, 621–624. coastal system. Estuar. Coast. Shelf Sci. 41, 475–490. Sarmiento J. L., Hughes T. M. C., Stouffer R. J., and Manabe S. Paasche E. (1999) Reduced calcite production under (1998) Simulated response of the ocean carbon cycle to light-limited growth: a comparative study of three clones of anthropogenic climate warming. Nature 393, 245–249. Emiliania huxleyi (Prymnesiophyceae). Phycologia 38, Scholten J. C., et al. (2001) Trapping efficiency of sediment 508–516. traps from the deep eastern North Atlantic: 230Th calibration. Pahlow M. and Riebesell U. (2000) Temporal trends in deep Deep-Sea Res. II 48, 243–268. ocean Redfield ratios. Science 287, 831–833. Schumann R. and Rentsch D. (1998) Staining particulate Passow U. (2000) Formation of transparent exopolymer organic matter with DTAF—a fluorescence dye for carbo- particles, TEP, from dissolved precursor material. Mar. hydrates and protein: a new approach and application of a Ecol. Prog. Ser. 192, 1–11. 2D image analysis system. Mar. Ecol. Prog. Ser. 163, Passow U. (2002) Transparent exopolymer particles (TEP) in 77–88. aquatic environments. Prog. Oceanogr. 55, 287–333. Shaffer G. (1996) Biogeochemical cycling in the global ocean: Passow U., Shipe R. F., Murray A., Pak D., Brzezinski M. A., 2. New production, Redfield ratios, and remineralization in and Alldredge A. L. (2001) The origin of transparent the organic pump. J. Geophys. Res. 101, 3723–3745. References 111

Shanks A. L. and Trent J. D. (1980) Marine snow: sinking rates Takahashi T., Broecker W. S., and Langer S. (1985) Redfield and potential role in vertical flux. Deep-Sea Res. 27A, ratio based on chemical data from surfaces. 137–143. J. Geophys. Res. 90, 6907–6924. Shemesh A., Mortlock R. A., Smith R. J., and Froelich P. N. Takeda S. (1998) Influence of iron availability on nutrient (1988) Determination of Ge/Si in marine siliceous micro- consumption ratio of diatoms in oceanic waters. Nature 393, : separation, cleaning and dissolution of diatoms and 774–777. . Mar. Chem. 25, 305–323. Tappan H. and Loeblich A. R., Jr. (1973) Evolution of the Shulenberger E. and Reid J. L. (1981) The Pacific shallow oceanic plankton. Earth Sci. Rev. 9, 207–240. oxygen maximum, deep chlorophyll maximum, and primary Teira E., Pazo´ M. J., Serret P., and Ferna´ndez E. (2001) productivity, reconsidered. Deep-Sea Res. 28A, 901–919. Dissolved organic carbon production by microbial popu- Shuskina E. A. (1985) Production of principal ecological lations in the Atlantic Ocean. Limnol. Oceanogr. 46, groups of plankton in the epipelagic zone of the ocean. 1370–1377. Oceanology 25, 653–658. Timmermans K. R., Snoek J., Gerringa L. J. A., Zondervan I., Siever R. (1991) Silica in the Oceans: biological–geochemical and de Baar H. J. W. (2001) Not all eukaryotic algae can interplay. In Scientists on Gaia (eds. S. Schneider and P. replace zinc with cobalt: Chaetoceros calcitrans (Bacillar- Boston). MIT Press, pp. 287–295. iophyceae) versus Emiliania huxleyi (Prymnesiophyceae). Smayda T. J. (1970) The suspension and sinking of Limnol. Oceanogr. 46, 699–703. phytoplankton in the sea. Oceanogr. Mar. Biol. Ann. Rev. Tortell P. D., Reinfelder J. R., and Morel F. M. M. (1997) 8, 353–414. Active uptake of bicarbonate by diatoms. Nature 390, Smith D. C., Simon M., Alldredge A. L., and Azam F. (1992) 243–244. Intense hydrolytic enzyme activity on marine aggregates and Tre´guer P., Nelson D. M., Van Bennekom A. J., DeMaster D. J., implications for rapid particle dissolution. Nature 359, Leynaert A., and Que´guiner B. (1995) The silica balance in 139–142. the world ocean: a reestimate. Science 268, 375–379. Smith D. C., Steward G. F., Long R. A., and Azam F. (1995) Turner J. T. (2002) Zooplankton fecal pellets, marine snow and Bacterial mediation of carbon fluxes during a diatom bloom sinking phytoplankton blooms. Aquat. Microbiol. Ecol. 27, in a mesocosm. Deep-Sea Res. II 42, 75–97. 57–102. Soendergaard M., Williams P. J. L., Cauwet G., Riemann B., Tyrrell T. (1999) The relative influences of nitrogen and Robinson C., Terzic S., Woodward E. M. S., and Worm J. phosphorus on oceanic primary production. Nature 400, (2000) Net accumulation and flux of dissolved organic 525–531. carbon and dissolved organic nitrogen in marine plankton Verity P. G., Williams S. C., and Hong Y. (2000) Formation, communities. Limnol. Oceanogr. 45, 1097–1111. degradation, and mass: volume ratios of detritus derived Steemann Nielsen E. (1952) The use of radioactive carbon from decaying phytoplankton. Mar. Ecol. Prog. Ser. 207, (14C) for measuring organic production in the sea. J. Cons. 53–68. Int. Explor. Mer. 18, 117–140. Ware J. R., Smith S. V., and Reaka-Kudla M. L. (1992) Steinberg D. K. (1995) Diet of (Scolpalatum vorax) reefs: sources or sinks of atmospheric CO2. Coral Reefs 11, associated with mesopelagic detritus (giant larvacean 127–130. houses) in Monterey Bay, California Mar. Biol. 122, Watson A. J., et al. (1994) Minimal effect of iron fertilization 571–584. on sea-surface carbon dioxide concentrations. Nature 371, Steinberg D. K., Pilskaln C. H., and Silver M. W. (1998) 143–145. Contribution of zooplankton associated with detritus to Watson A. J., Bakker D. C. E., Ridgwell A. J., Poyd P. W., and sediment trap “swimmer” carbon in Monterey Bay, Califor- Law C. S. (2000) Effect of iron supply on Southern Ocean nia, USA. Mar. Ecol. Prog. Ser. 164, 157–166. CO2 uptake and implications for glacial atmospheric CO2. Steinberg D. K., Carlson C. A., Bates N. R., Johnson R. J., Nature 407, 730–733. Michaels A. F., and Knap A. P. (2001) Overview of the US Weiss R. F., Broecker W. S., Craig H., and Spencer D. (1983) JGOFS Bermuda Atlantic Time-series Study (BATS): a GEOSECS Indian Ocean Expedition: Volume 5. Hydro- decade-scale look at ocean biology and biogeochemistry. graphic Data 1977–1978. National Science Foundation. Deep-Sea Res. II 48, 1405–1447. Williams P. M. and Druffel E. R. M. (1987) Radiocarbon in Strass V. H. (2002) EisenEx-1: test of the iron hypothesis in a dissolved organic matter in the central north Pacific Ocean. Southern Ocean eddy. EOS: Trans., AGU Abstracts of 2002 Nature 330, 246–248. Ocean Sciences Meeting. Wolf-Gladrow D. A., Riebesell U., Burkhardt S., and Bijma J. Strickland J. D. H. and Parsons T. R. (1968) A Practical (1999) Direct effects of CO2 concentration on growth and Handbook of Seawater Analysis. Research Board isotopic composition of marine plankton. Tellus (B Chem. of Canada. Phys. Meteorol.) 51B, 176–461. Strom S. L., Benner R., Ziegler S., and Dagg M. J. (1997) Zeebe R. E. and Wolf-Gladrow D. (2001) CO2 in Seawater: Planktonic grazers are a potentially important source of Equilibrium, Kinetics, Isotopes. Elsevier. marine dissolved organic carbon. Limnol. Oceanogr. 42, Zehr J. P. and Ward B. B. (2002) Nitrogen cycling in the ocean: 1364–1374. new perpsectives on processes and paradigms. Appl. Suess E. (1980) Particulate organic carbon flux in the oceans- Environ. Microbiol. 68, 1015–1024. surface productivity and oxygen utilization. Nature 288, Zehr J. P., Waterbury J. B., Turner P. J., Montoya J. P., 260–263. Omoregie E., Steward G. F., Hansen A., and Karl D. M. Sunda W. G. and Huntsman S. A. (1995a) Iron uptake and (2001) Unicellular cyanobacteria fix N2 in the subtropical growth limitation in oceanic and coastal phytoplankton. North Pacific Ocean. Nature 412, 635–638. Mar. Chem. 50, 189–206. Zhang J. and Nozaki Y. (1996) Rare earth elements and yttrium Sunda W. G. and Huntsman S. A. (1995b) Cobalt and zinc in seawater: ICP_MS determinations in the East Caroline, interreplacement in marinephytoplankton: biological and Coral Sea, and Southern Fiji basins of the western Pacific geochemical implications. Limnol. Oceanogr. 40, Ocean. Geochim. Cosmochim. Acta 60, 4631–4644. 1404–1417. Zhuang G., Duce R. A., and Kester D. R. (1990) The Sundquist E. T. (1993) The global carbon dioxide budget. dissolution of atmospheric iron in surface seawater of the Science 259, 934–941. open ocean. J. Geophys. Res. 95, 16207–16216. q 2003, Elsevier Ltd. All rights reserved Treatise on Geochemistry No part of this publication may be reproduced, stored in a retrieval system or ISBN (set): 0-08-043751-6 transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Volume 6; (ISBN: 0-08-044341-9); pp. 83–111