8.05 Biogeochemistry of Primary Production in the Sea P. G. Falkowski Rutgers University, New Brunswick, NJ, USA

8.05.1 INTRODUCTION 186 8.05.1.1 The Two Carbon Cycles 186 8.05.1.2 A Primer on Redox Chemistry 186 8.05.2 CHEMOAUTOTROPHY 187 8.05.3 PHOTOAUTOTROPHY 188 8.05.3.1 Selective Forces in the Evolution of Photoautotrophy 188 8.05.3.2 Selective Pressure in the Evolution of Oxygenic Photosynthesis 189 8.05.4 PRIMARY PRODUCTIVITY BY PHOTOAUTOTROPHS 189 8.05.4.1 What are Photoautotrophs? 189 8.05.4.1.1 The red and green lineages 191 8.05.4.2 Estimating Chlorophyll Biomass 191 8.05.4.2.1 Satellite based algorithms for ocean color retrievals 193 8.05.4.3 Estimating Net Primary Production 194 8.05.4.3.1 Global models of net primary production for the ocean 194 8.05.4.4 Quantum Efficiency of NPP 195 8.05.4.4.1 Comparing efficiencies for oceanic and terrestrial primary production 196 8.05.5 EXPORT, NEW AND “TRUE NEW” PRODUCTION 196 8.05.5.1 Steady-state versus Transient State 197 8.05.6 NUTRIENT FLUXES 198 8.05.6.1 The Redfield Ratio 198 8.05.7 NITRIFICATION 198 8.05.7.1 Carbon Burial 199 8.05.7.2 Carbon Isotope Fractionation in Organic Matter and Carbonates 199 8.05.7.3 Balance between Net Primary Production and Losses 199 8.05.7.4 Carbon Burial in the Contemporary Ocean 201 8.05.7.5 Carbon Burial in the Precambrian Ocean 201 8.05.8 LIMITING MACRONUTRIENTS 201 8.05.8.1 The Two Concepts of Limitation 201 8.05.9 THE EVOLUTION OF THE NITROGEN CYCLE 202 8.05.10 FUNCTIONAL GROUPS 203 8.05.10.1 Siliceous Organisms 204 8.05.10.2 Calcium Carbonate Precipitation 204 8.05.10.3 Vacuoles 205 8.05.11 HIGH-NUTRIENT, LOW-CHLOROPHYLL REGIONS—IRON LIMITATION 206

8.05.12 GLACIAL–INTERGLACIAL CHANGES IN THE BIOLOGICAL CO2 PUMP 207 8.05.13 IRON STIMULATION OF NUTRIENT UTILIZATION 207

8.05.14 LINKING IRON TO N2 FIXATION 207 8.05.15 OTHER TRACE-ELEMENT CONTROLS ON NPP 208 8.05.16 CONCLUDING REMARKS 209 ACKNOWLEDGMENTS 209 REFERENCES 210

185 186 Biogeochemistry of Primary Production in the Sea As the present condition of nations is the result of carbon to form organic matter, the overwhelming many antecedent changes, some extremely remote majority of which is oxidized back to inorganic and others recent, some gradual, others sudden and carbon by respiratory metabolism (Schlesinger, violent, so the state of the natural world is the result of 1997). This cycle, which is observable on time- a long succession of events, and if we would enlarge scales of days to millenia, is driven by reduction— our experience of the present economy of nature, we must investigate the effects of her operations in oxidation (redox) reactions that evolved over former epochs. ,2 Gyr, first in microbes, and subsequently in multicellular organisms (Falkowski et al., 1998). Charles Lyell, A very small fraction of the reduced carbon Principles of Geology, 1830 escapes respiration and becomes incorporated into the lithosphere. In so doing, some of the organic matter is transferred to the slow carbon 8.05.1 INTRODUCTION cycle. In this chapter, we will focus primarily on this fast, biologically mediated carbon cycle in Earth is the only planet in our solar system that the sea, and the supporting biogeochemical contains vast amounts of liquid water on its processes and feedbacks. surface and high concentrations of free molecular oxygen in its atmosphere. These two features are not coincidental. All of the original oxygen on 8.05.1.2 A Primer on Redox Chemistry Earth arose from the photobiologically catalyzed splitting of water by unicellular photosynthetic The biologically mediated redox reactions organisms that have inhabited the oceans for at cycle carbon through three mobile pools: the least 3 Gyr. Over that period, these atmosphere, the ocean, and the biosphere. Of organisms have used the hydrogen atoms from these, the ocean is by far the largest (Table 1); water and other substrates to form organic matter however, more than 98% of this carbon is found in its oxidized state as CO2 and its hydrated from CO2 and its hydrated equivalents. This 2 22 process, the de novo formation of organic matter equivalents, HCO3 and CO3 . To form organic from inorganic carbon, or primary production, is molecules, the inorganic carbon must be chemi- the basis for all life on Earth. In this cally reduced, a process that requires the addition chapter, we examine the evolution and biogeo- of hydrogen atoms (not just protons, but protons chemical consequences of primary production in plus electrons) to the carbon atoms. Broadly the sea and its relationship to other biogeochem- speaking, these biologically catalyzed reduction ical cycles on Earth. reactions are carried out by two groups of organisms, chemoautotrophs and photoauto- trophs, which are collectively called primary

8.05.1.1 The Two Carbon Cycles Table 1 Carbon pools in the major reservoirs on Earth. There are two major carbon cycles on Earth. The two cycles operate in parallel. One cycle is Pools Quantity £ 15 slow and abiotic. Its effects are observed on ( 10 g) multimillion-year timescales and are dictated by Atmosphere 720 tectonics and weathering (Berner, 1990). In this Oceans 38,400 cycle, CO2 is released from the mantle to the atmosphere and oceans via vulcanism and seafloor Total inorganic 37,400 Surface layer 670 spreading, and removed from the atmosphere and Deep layer 36,730 ocean primarily by reaction with silicates to form Dissolved organic 600 carbonates in the latter reservoir. Most of the carbonates are subsequently subducted into the Lithosphere mantle, where they are heated, and their carbon is Sedimentary carbonates .60,000,000 Kerogens 15,000,000 released as CO2 to the atmosphere and ocean, to carry out the cycle again. The chemistry of this Terrestrial biosphere (total) 2,000 cycle is dependent on acid–base reactions, and Living biomass 600–1,000 would operate whether or not there was life on the Dead biomass 1,200 planet (Kasting et al., 1988). This slow carbon Aquatic biosphere 1–2 cycle is a critical determinate of the concentration Fossil fuels 4,130 of CO2 in Earth’s atmosphere and oceans on timescales of tens and hundreds of millions of Coal 3,510 years (Kasting, 1993). Oil 230 Gas 140 The second carbon cycle is dependent on the Other (peat) 250 biologically catalyzed reduction of inorganic Chemoautotrophy 187 producers. The organic carbon they synthesize 8.05.2 CHEMOAUTOTROPHY fuels the growth and respiratory demands of the primary producers themselves and all remaining Organisms capable of reducing sufficient inor- organisms in the ecosystem. ganic carbon to grow and reproduce in the dark All redox reactions are coupled sequences. without an external organic carbon source are Reduction is accomplished by the addition of an called chemoautotrophs (literally, “chemical self- electron or hydrogen atom to an atom or molecule. feeders”). Genetic analyses suggest that chemo- In the process of donating an electron to an autotrophy evolved very early in Earth’s history, acceptor, the donor molecule is oxidized. Hence, and is carried out exclusively by prokaryotic redox reactions require pairs of substrates, and organisms in both the Archea and Bacteria can be described by a pair of partial reactions, or superkingdoms (Figure 1). half-cells: Early in Earth’s history, the biological reduction of inorganic carbon may have been þ ð 2Þ$ ð Þ Aox n e Are 1a directly coupled to the oxidation of H2. At present, however, free H2 is scarce on the planet’s surface. 2 Bred 2 nðe Þ$Box ð1bÞ Rather, most of the hydrogen on the surface of Earth is combined with other atoms, such as sulfur The tendency for a molecule to accept or release or oxygen. Activation energy is required to break an electron is therefore “relative” to some other these bonds in order to extract the hydrogen. One molecule being capable of conversely releasing source of energy is chemical bond energy itself. or binding an electron. Chemists scale this For example, the ventilation of reduced mantle tendency, called the redox potential, E; relative gases along tectonic plate subduction zones on to the reaction the seafloor provides hydrogen in the form of H2S. Several types of microbes can couple the þ 2 H2 $ 2H þ 2e ð2Þ oxidation of H2S to the reduction of inorganic carbon, thereby forming organic matter in the which is arbitrarily assigned an E of 0 at pH 0, and absence of light. : is designated E0 Biologists define the redox Ultimately all chemoautotrophs depend on a potential at pH 7, 298 K (i.e., room temperature) nonequilibrium redox gradient, without which and 1 atm pressure (¼101.3 kPa). When so there is no thermodynamic driver for carbon defined, the redox potential is denoted by the fixation. For example, the reaction involving the 0 : 0 symbols E0 or sometimes Em7 The E0 for a oxidation of H2S by microbes in deep-sea vents standard hydrogen electrode is 2420 mV. described above is ultimately coupled to oxygen

Three domains of life Bacteria Eukarya

Proteobacteria Archea Animals Slime Molds Gram-positive bacteria Entamoebae Green non-sulfur bacteria Halobacterium Methanobacterium Fungi Cyanobacteria Plants Methanococcus Thermoplasma Flamobacteria Thermoproteus Thermococcus Ciliates Suiflobus Archeo globus Methanopyrus Euglena Thermotogz Pyrodictium Microsporidia Aquifex Diplomanads

Universal Pathways of autotrophic CO fixation Ancestor 2 Reductive citric acid cycle Reductive acetyl CoA pathway Reductive hydroxypropionante pathway Cathin-Bonion cycle

Figure 1 The distribution of autotrophic metabolic pathways among taxa within the three major domains of life (as inferred from 16S ribosomal RNA sequences (Pace, 1997)). Specific metabolic pathways are indicated. 188 Biogeochemistry of Primary Production in the Sea in the ocean interior. Hence, this reaction is depen- Moreover, magma chambers, vulcanism and vent dent on the chemical redox gradient between the fluid fluxes are tied to tectonic subduction and ventilating mantle plume and the ocean interior spreading regions, which are transient features of that thermodynamically favors oxidation of the Earth’s crust and hence only temporary habitats plume gases. Maintaining such a gradient requires for chemoautotrophs. In the Archean and early a supply of energy, either externally, from Proterozoic oceans, the chemoautotrophs would radiation (solar or otherwise), or internally, via have already been dispersed throughout the oceans planetary heat and tectonics, or both. by physical mixing and helping to colonize new The overall contribution of chemoautotrophy in vent regions. This same dispersion process would thecontemporaryoceantotheformationof have also helped ancestral chemoautotrophs organic matter is relatively small, accounting for exploit solar energy near the ocean surface. ,1% of the total annual primary production in the Although the processes that selected the photo- sea. However, this process is critical in coupling synthetic reactions as the major energy transduc- reduction of carbon to the oxidation of low-energy tion pathway remain obscure, central hypotheses substrates, and is essential for completion of have emerged based on our understanding of the several biogeochemical cycles. evolution of Earth’s carbon cycle, the evolution of photosynthesis, biophysics, and molecular phylo- geny. Photoautotrophs are found in all three major 8.05.3 PHOTOAUTOTROPHY superkingdoms (Figure 1); however, there are The oxidation state of the ocean interior is a very few known Archea capable of this form of consequence of a second energy source: light, metabolism. Efficient photosynthesis requires which drives photosynthesis. Photosynthesis is a harvesting solar radiation, and hence the evolution redox reaction of the general form: of a light harvesting system. While some Archea and Bacteria use the pigment-protein rhodopsin, 2H2A þ CO2 þ light ! ðCH2OÞþH2O þ 2A by far, the most efficient and ubiquitous light ð3Þ harvesting systems are based on chlorins. The metabolic pathway for the synthesis of porphyrins where A is an atom, e.g., S. In this formulation, and chlorins is one of the oldest in biological light is specified as a substrate, and a fraction of evolution, and is found in all chemoautotrophs the light energy is stored as chemical bond energy (Xiong et al., 2000). Mulkidjanian and Junge in the organic matter. Organisms capable of (1997) proposed that the chlorin-based photosyn- reducing inorganic carbon to organic matter by thetic energy conversion apparatus originally using light energy to derive the source of reductant arose from the need to prevent UV radiation or energy are called photoautotrophs. Analyses of from damaging essential macromolecules such as genes and metabolic sequences strongly suggest nucleic acids and proteins. The UV excitation that the machinery for capturing and utilizing light energy could be transferred from the aromatic as a source of energy to extract reductants was amino acid residues in the macromolecule to a built on the foundation of chemoautotrophic blue absorption band of membrane-bound chlorins carbon fixation; i.e., the predecessors of photo- to produce a second excited state which sub- autotrophs were chemoautotrophs. The evolution sequently decays to the lower-energy excited of a photosynthetic process in a chemoauto- singlet. This energy dissipation pathway can be troph forces consideration of both the selective harnessed to metabolism if the photochemically forces responsible (why) and the mechanism of produced, charge-separated, primary products are evolution (how). prevented from undergoing a back-reaction, but rather form a biochemically stable intermediate 8.05.3.1 Selective Forces in the Evolution reductant. This metabolic strategy was selected of Photoautotrophy for the photosynthetic reduction of CO2 to carbohydrates, using reductants such as S22 or Reductants for chemoautotrophs are generally Fe2þ, which have redox potentials that are too deep in the Earth’s crust. Vent fluids are produced positive to reduce CO2 directly. in magma chambers connected to the Atheno- The synthesis of reduced (i.e., organic) carbon sphere. As such, the supply of vent fluids is and the oxidized form of the electron donor virtually unlimited. While the chemical disequili- permits a photoautotroph to use “respiratory” bria between vent fluids and bulk seawater metabolism, but operate them in reverse. However, provides a sufficient thermodynamic gradient to not all of the reduced carbon and oxidants remain continuously support chemoautotrophic meta- accessible to the photoautotrophs. In the oceans, bolism in the contemporary ocean, in the early cells tend to sink, carrying with them organic Earth the oceans would not have had a suffi- carbon. The oxidation of Fe2þ forms insoluble ciently large thermodynamic energy potential to Fe3þ salts that precipitate. The sedimentation and support a pandemic outbreak of chemoautotrophy. subsequent burial of organic carbon and Fe3þ Primary Productivity by Photoautotrophs 189 removes these components from the water column. quartet of manganese atoms, that sequentially Without replenishment, the essential reductants extract electrons, one at a time, from 2H2O mole- for anoxygenic photosynthesis would eventually cules, releasing gaseous O2 to the environment, become depleted in the surface waters. Thus, the and storing the reductants on biochemical necessity to regenerate reductants potentially intermediates. prevented anoxygenic photoautotrophs from pro- The photochemically produced reductants gene- viding the major source of fixed carbon on Earth rated by the reactions schematically outlined in for eternity. Major net accumulation of reduced Equation (5) are subsequently used in the fixation organic carbon in Proterozoic sediments implies (fixation is an archaic term meaning to make non- local depletion of reductants such as S22 and Fe2þ volatile, as in the chemical conversion of a gas to a from the euphotic zone of the ocean. These solid phase) of CO2 by a suite of enzymes that can limitations almost certainly provided the evolu- operate in vitro in darkness and, hence, the tionary selection pressure for an alternative ensemble of these reactions are called the dark electron donor. reactions. At pH 7 and 25 8C, the formation of glucose from CO2 requires an investment of 915 cal mol21. If water is the source of reductant, the overall efficiency for photosynthetic reduction 8.05.3.2 Selective Pressure in the Evolution of CO to glucose is ,30%; i.e., 30% of the of Oxygenic Photosynthesis 2 absorbed solar radiation is stored in the chemical

H2O is a potentially useful biological reductant bonds of glucose molecules. with a vast supply on Earth relative to any redox- active solute dissolved in it. Liquid water contains ,100 kmol of H atoms per m3,and,given .1018 m3 of water in the hydrosphere and cryo- 20 8.05.4 PRIMARY PRODUCTIVITY sphere, .10 kmol of reductant are potentially BY PHOTOAUTOTROPHS accessible. Use of H2O as a reductant for CO2, however, requires a larger energy input than does When we subtract the costs of all other the use of Fe2þ or S22. Indeed, to split water by metabolic processes by the chemoautotrophs and light energy requires 0.82 eV at pH 7 and photoautotrophs, the organic carbon that remains 298 K. Utilizing light at such high energy levels is available for the growth and metabolic costs of required the evolution of a new photosynthetic heterotrophs. This remaining carbon is called net pigment, chlorophyll a, which has a red (lowest primary production (NPP) (Lindeman, 1942). singlet) absorption band that is 200–300 nm blue From biogeochemical and ecological perspec- shifted relative to bacteriochlorophylls. Moreover, tives, NPP provides an upper bound for all other stabilization of the primary electron acceptor to metabolic demands in an ecosystem. If NPP is prevent a back-reaction necessitates thermody- greater than all respiratory consumption of the namic inefficiency that ultimately requires two ecosystem, the ecosystem is said to be net light-driven reactions operating in series. This seq- autotrophic. Conversely, if NPP is less than all uential action of two photochemical reactions is respiratory consumption, the system must either unique to oxygenic photoautotrophs and presum- import organic matter from outside its bounds, or ably involved horizontal gene transfer through one it will slowly run down—it is net heterotrophic. or more symbiotic events (Blankenship, 1992). It should be noted that NPP and photosynthesis In all oxygenic photoautotrophs, Equation (3) are not synonymous. On a planetary scale, the can be modified to: former includes chemoautrophy, the latter does Chl a not. Moreover, photosynthesis per se does not 2H2O þ CO2 þ light ! ðCH2OÞþH2O þ O2 include the integrated respiratory term for the ð4Þ photoautotrophs themselves (Williams, 1993). In where Chl a is the pigment chlorophyll a reality, that term is extremely difficult to measure exclusively utilized in the reaction. Equation (4) directly, hence NPP is generally approximated implies that somehow chlorophyll a catalyzes a from measurements of photosynthetic rates reaction or a series of reactions whereby light integrated over some appropriate length of time energy is used to oxidize water: (a day, month, season, or a year) and respiratory costs are either assumed or neglected. Chl a þ 2H2O þ light ! 4H þ 4e þ O2 ð5Þ yielding gaseous, molecular oxygen. Hidden 8.05.4.1 What are Photoautotrophs? within Equation (5) are complex suites of biological innovations that have heretofore not In the oceans, oxygenic photoautotrophs are been successfully mimicked in vitro by humans. a taxonomically diverse group of mostly At the core of the water splitting complex is a single-celled, photosynthetic organisms that drift 190 Biogeochemistry of Primary Production in the Sea with currents. In the contemporary ocean, these By comparison, higher plants are comprised of organisms, called phytoplankton (derived from .2.5 £ 105 species, almost all of which are Greek, meaning to wander), are comprised of contained within one class in one division. Thus, ,2 £ 104 species distributed among at least unlike terrestrial plants, phytoplankton are repre- eight taxonomic divisions or phyla (Table 2). sented by relatively few species but they are

Table 2 The taxonomic classification and species abundances of oxygenic photosynthetic organisms in aquatic and terrestrial ecosystems. Note that terrestrial ecosystems are dominated by relatively few taxa that are species rich, while aquatic ecosystems contain many taxa but are relatively species poor. Taxonomic group Known species Marine Freshwater

Empire: Bacteria (¼Prokaryota) Kingdom: Eubacteria Subdivision: Cyanobacteria (sensu strictu) 1,500 150 1,350 (¼Cyanophytes, blue-green algae) Subdivision: Chloroxybacteria 321 (¼Prochlorophyceae) Empire: Eukaryota Kingdom: Protozoa Division: Euglenophyta 1,050 30 1,020 Class: Euglenophyceae Division: Dinophyta (Dinoflagellates) Class: Dinophyceae 2,000 1,800 200 Kingdom: Plantae Subkingdom: Biliphyta Division: Glaucocystophyta Class: Glaucocystophyceae 13 Division: Rhodophyta Class: Rhodophyceae 6,000 5,880 120 Subkingdom: Viridiplantae Division: Chlorophyta Class: Chlorophyta 2,500 100 2,400 Prasinophyceae 120 100 20 Ulvophyceae 1,100 1,000 100 Charophyceae 12,500 100 12,400 Division: Bryophyta (mosses, liverworts) 22,000 1,000 Division: Lycopsida 1,228 70 Division: Filicopsida (ferns) 8,400 94 Division: Magnoliophyta (flowering plants) (240,000) Subdivision: Monocotyledoneae 52,000 55 455 Subdivision: Dicotyledoneae 188,000 391 Kingdom: Chromista Subkingdom: Chlorechnia Division: Chlorarachniophyta Class: Chlorarachniophyceae 3–4 3–4 0 Subkingdom: Euchromista Division: Crytophyta Class: Crytophyceae 200 100 100 Division: Haptophyta Class: Prymensiophyceae 500 100 400 Division: Heterokonta Class: Bacillariophyceae (diatoms) 10,000 5,000 5,000 Chrysophyceae 1,000 800 200 Eustigmatophyceae 12 6 6 Fucophyceae (brown algae) 1,500 1,497 3 Raphidophyceae 27 10 17 Synurophyceae 250 250 Tribophyceae (Xanthophyceae) 600 50 500 Kingdom: Fungi Division: Ascomycontina (lichens) 13,000 15 20

Source: Falkowski (1997). Primary Productivity by Photoautotrophs 191 phylogenetically diverse. This deep taxonomic All other oxygen-producing organisms in the diversity is reflected in their evolutionary history ocean are eukaryotic, i.e., they contain internal and ecological function (Falkowski, 1997). organelles, including a nucleus, one or more Within this diverse group of organisms, three chloroplasts, one or more mitochondria, and, in basic evolutionary lineages are discernable some cases, a membrane-bound storage compart- (Delwiche, 2000). The first contains all prokaryotic ment, the vacuole. Within the eukaryotes, we can oxygenic phytoplankton, which belong to one class distinguish two major groups, both of which of bacteria, namely, the cyanobacteria. Cyanobac- appear to have descended from a common teria are the only known oxygenic photoautotrophs ancestor thought to be the endosymbiotic appro- that existed prior to ,2.5 Gyr BP (Ga) (Lipps, priation of a cyanobacterium into a heterotrophic 1993; Summons et al., 1999). These prokaryotes host cell (Delwiche, 2000). The appropriated numerically dominate the photoautotrophic com- cyanobacterium became a chloroplast. munity in contemporary marine ecosystems and their continued success bespeaks an extraordinary adaptive capacity. At any moment in time, there are 8.05.4.1.1 The red and green lineages ,1024 cyanobacterial cells in the contemporary oceans. To put that into perspective, the number of In one group of eukaryotes, chlorophyll b was cyanobacterial cells in the oceans is two orders of synthesized as a secondary pigment; this group magnitude more than all the stars in the sky. forms the “green lineage,” from which all higher The evolutionary history of cyanobacteria is plants have descended. The green lineage played obscure. The first microfossils assigned to this a major role in oceanic food webs and the carbon group were identified in cherts from 3.1 Ga by cycle from ca. 1.6 Ga until the end-Permian Schopf (Schopf, 1993). Macroscopic stromabo- extinction, ,250 Ma (Lipps, 1993). Since that lites, which are generally of biological (oxygenic time however, a second group of eukaryotes has photoautotrophic) origin, are first found in strata a risen to ecological prominence in the oceans; that few hundred million years younger. However, group is commonly called the “red lineage” much of the fossil evidence provided by Schopf (Figure 2). The red lineage is comprised of several (e.g., Schopf, 1993) has been questioned (Brasier major phytoplankton divisions and classes, of et al., 2002), and many researchers believe in a later which the diatoms, dinoflagellates, haptophytes origin. The origin of this group is critical to (including the coccolithophorids), and the chryso- phytes are the most important. All of these groups establishing when net O2 production (and hence, an oxidized atmosphere) first occurred on the planet. are comparatively modern organisms; indeed, the rise of dinoflagellates and coccolithophorids Although photodissociation of H2O vapor could approximately parallels the rise of dinosaurs on have provided a source of atmospheric O2 in the land, while the rise of diatoms approximately Archean, the UV absorption cross-section of O2 constrains the reaction, and theoretical calculations parallels the rise of mammals in the Cenozoic. supported by geochemical evidence suggest that The burial and subsequent diagenesis of organic carbon produced primarily by members of the red prior to ca. 2.4 Ga atmospheric O2 was less than 1025 of the present level (Holland and Rye, 1998; lineage in shallow seas in the Mesozoic era Pavlov and Kasting, 2002). There was a lag provide the carbon source for many of the between the first occurrence of oxygenic photo- petroleum reservoirs that have been exploited for the past century by humans. synthesis and a global buildup of O2 possibly due to the presence of alternative electron acceptors, especially Fe2þ and S22 in the ocean. Indeed, the dating of oxidation of Earth’s oceans and atmos- 8.05.4.2 Estimating Chlorophyll Biomass phere is, in large measure, based on analysis of the chemical precipitation of oxidized iron As implied in Equation (4), given an abundance in sedimentary rocks (the “Great Rust Event of the two physical substrates, CO2 and H2O, (Holland and Rye, 1998) and the mass-independent primary production is, to first order, dependent on (Farquhar et al., 2002) and mass-dependent the concentration of the catalyst Chl a and light. (Habicht et al., 2002) fractionation of sulfur The distribution of Chl a in the upper ocean can be isotopes. The ensemble of these analyses indicate discerned from satellite images of ocean color. that atmospheric oxygen rose sharply, from The physical basis of the measurement is virtually insignificant levels, to between 1% and straightforward; making the measurements is 10% of the present atmospheric concentration over technically challenging. Imagine two small par- a 100 Myr period beginning ca. 2.4 Ga. Thus, there cels of water that are adjacent to each other. As may be as much as a 1 Gyr or as little as a 100 Myr photons from the sun enter the water column, they gap between the origin of the first oxygenic are either absorbed or scattered. Water itself photoautotrophs and oxygenation of Earth’s absorbs red wavelengths of light, at shorter oceans and atmosphere. wavelengths of the visible spectrum, light is not 192 Biogeochemistry of Primary Production in the Sea

Figure 2 The basic pathway leading the evolution of eukaryotic algae. The primary symbiosis of a cyanobacterium with a apoplastidic host gave rise to both chlorophyte algae and red algae. The chlorophyte line, through secondary symbioses, gave rise to the “green” line of algae, one division of which was the predecessor of all higher plants. Secondary symbioses in the red line with various host cells gave rise to all the chromophytes, including diatoms, cryptophytes, and haptophytes (after Delwiche, 2000). as efficiently absorbed. However, because water downwelling light stream. The incoherence, in molecules can randomly move from one adjacent turn, increases the probability of photon scattering parcel to another, there are continuous minor (a process called “fluctuation density scattering”), changes in density and hence in the refractive such that light in the shorter wavelengths is more index of the water parcels. These minor changes in likely to be scattered back to space (Einstein, refractive index lead to incoherence in the 1910; Morel, 1974). If the ocean contained sterile, Primary Productivity by Photoautotrophs 193 pure seawater, an observer looking at the surface One limitation of satellite images of ocean from space would see the oceans as blue. chlorophyll is that they do not provide information However, chlorophyll a has a prominent absorp- about the vertical distribution of phytoplankton. tion band in the blue portion of the spectrum. The water-leaving radiances visible to an observer Hence, in the presence of chlorophyll some of the outside of the ocean are confined to the upper 20% downwelling and upwelling photons from the sun of the euphotic zone (which is empirically defined are absorbed by the phytoplankton themselves as the depth to which 1% of the solar radiation and the ocean becomes optically darker. As penetrates). In the open ocean there is almost chlorophyll concentrations increase even further, always a subsurface chlorophyll maximum that is blue wavelengths are largely eliminated from not visible to satellite ocean color sensors. A the outbound reflectance, and the ocean appears number of numerical models have been developed optically dark green (Morel, 1988). to estimate the vertical distribution of chlorophyll based on satellite color data (Berthon and Morel, 8.05.4.2.1 Satellite based algorithms for 1992; Platt, 1986). The models rely on statistical ocean color retrievals parametrizations and require numerous in situ observations to obtain typical profiles for a given Empirically, satellite sensors that measure area of the world ocean (Morel and Andre, 1991; ocean color utilize a number of wavelengths. In Platt and Sathyendranath, 1988). In addition, large addition to the blue and green region, red and far- quantities of phytoplankton associated with the red spectra are determined to derive corrections bottom of ice flows in both the Arctic and for scattering and absorption of the outbound or Antarctic are not visible to satellite sensors but reflected radiation from the ocean by the atmos- do contribute significantly to the primary pro- phere. In fact, only a very small fraction (,5%) of duction in the polar seas (Smith and Nelson, the light leaving the ocean is observed by a 1990). Despite these deficiencies, the satellite data satellite; the vast majority of the photons are allow high-resolution, synoptic observations of scattered or absorbed in the atmosphere. How- the temporal and spatial changes in phytoplankton ever, based on the ratio of blue-green light that is chlorophyll in relation to the physical circulation reflected from the ocean, estimates of photosyn- of the atmosphere and ocean on a global scale. thetic pigments are derived. It should be pointed The global distribution of phytoplankton out that the blue-absorbing region of the spectrum chlorophyll in the upper ocean for winter and is highly congested; it is virtually impossible to summer, derived from a compilation of satellite derive the fraction of absorption due solely to images, is shown in Figure 3. To a first order, the chlorophyll a as opposed to other photosynthetic images reveal how the horizontal and temporal pigments that absorb blue light. The estimation of distribution of phytoplankton is related to the chlorophyll a is based on empirical regression of physical circulation of the oceans, especially the the concentration of the pigment to the total blue- major features of the basin-scale gyres. For absorbing pigments (Gordon and Morel, 1983). example, throughout most of the central ocean Water-leaving radiances ðLWÞ at specific wave- basins, between 308 N and 308 S, phytoplankton lengths are corrected for atmospheric scattering biomass is extremely low, averaging 0.1–0.2 mg and absorption, and the concentration of chloro- chlorophyll a m23 at the sea surface. In these phyll is calculated from the ratios of blue and regions the vertical flux of nutrients is generally green light reflected from the water body. The extremely low, limited by eddy diffusion through calibration of the sensors is empirical and the thermocline. Most of the chlorophyll biomass specifically derived for individual satellites. is associated with the thermocline. Because there Examples of such algorithms for five satellites is no seasonal convective overturn in this are given in Table 3. latitude band, there is no seasonal variation in

Table 3 Algorithms used to calculate chlorophyll a (C) from remote sensing reflectance. R is determined as the maximum of the values shown. Sensor algorithms are for Sea-viewing Wide Field of view Sensor (SeaWiFS), Ocean Color and Temperature Scanner (OCTS), Moderate Resolution Imaging Spectroradiometer (MODIS), Coastal Zone Color Scanner (CZCS), and Medium Resolution Imaging Spectrometer (MERIS). Sensor Equation R

2 3 SeaWiFS/OC2 C ¼10:0ð0:34123:001Rþ2:811R 22:041R Þ 2 0:04 490=555 2 3 4 OCTS/OC4O C ¼10:0ð0::40522:900Rþ1:690R 20:530R 21:144R Þ 443.490.520=565 2 3 4 MODIS/OC3M C ¼10:0ð0:283022:753Rþ1:457R 20:659R 21:403R Þ 443.490=550 2 3 4 CZCS/OC3C C ¼10:0ð0:36224:066Rþ5:125R 22:645R 20:597R Þ 443.520=550 2 3 4 MERIS/OC4E C ¼10:0ð0:36822:814Rþ1:456R þ0:768R 21:292R Þ 443.490.510=560 2 3 4 SeaWiFS/OC4v4 C ¼10:0ð0:36623:067Rþ1:930R þ0:649R 21:532R Þ 443.490.510=555 2 3 SeaWiFS/OC2v4 C ¼10:0ð0:31922:336Rþ0:879R 20:135R Þ 2 0:071 490=555 194 Biogeochemistry of Primary Production in the Sea

Figure 3 Composite global images for winter and summer of upper-ocean chlorophyll concentrations (top panels) derived from satellite-based observations of ocean color, net primary production (middle panels) calculated based on the algorithms of Behrenfeld and Falkowski (1997a), and export production (lower panels) calculated from the model of Laws et al. (2000). phytoplankton chlorophyll. The chlorophyll con- bloom is associated with deep vertical convective centrations are slightly increased at the equator in mixing, which allows resupply of nutrients to the the Pacific and Atlantic Oceans, and south of the upper mixed layer of the ocean. This phenomenon equator in the Indian Ocean. In the equatorial does not occur in the Pacific due to a stronger regions the thermocline shoals laterally as a result vertical density gradient in that basin (driven by the of long-range wind stress at the surface (Pickard hydrological cycle). The North Atlantic bloom and Emery, 1990). The wind effectively piles up leads to a flux of organic matter into the ocean water along its fetch, thereby inclining the upper interior that is observed even at the seafloor. mixed layer. This results in increased nutrient In the southern hemisphere, phytoplankton fluxes, shallower mixed layers, and higher chloro- chlorophyll is generally lower at latitudes phyll concentrations on the eastern end of the symmetrical with the northern hemisphere in equatorial band, and decreased nutrient fluxes, the corresponding austral seasons. For example, deeper mixed layers, and lower chlorophyll in the austral summer (January–March), phyto- concentrations on the western end. This effect plankton chlorophyll is slightly lower between is most pronounced in the Pacific. The displace- 308 S and the Antarctic ice sheets than in the ment of the band south of the equator in the northern hemisphere in July to September Indian Ocean is primarily a consequence of basin (Yoder et al., 1993). scale topography. At latitudes above ,308, a seasonal cycle in chlorophyll can occur (Figure 3). In the northern 8.05.4.3 Estimating Net Primary Production hemisphere, areas of high chlorophyll are found 8.05.4.3.1 Global models of net primary in the open ocean of the North Atlantic in the spring production for the ocean and summer. The southern extent and intensity of the North Atlantic phytoplankton bloom are not Using satellite data to estimate upper- found in the North Pacific. The North Atlantic ocean chlorophyll concentrations, satellite-based Primary Productivity by Photoautotrophs 195 observations of incident solar radiation, atlases of This general model can be both expanded seasonally averaged sea-surface temperature, and (differentiated) and collapsed (integrated) with models that incorporate a temperature response respect to time and irradiance; however, the global function for photosynthesis, it is possible to results are fundamentally similar (Behrenfeld and estimate global net photosynthesis in the world Falkowski, 1997). The models predict that NPP in oceans (Antoine and Morel, 1996; Behrenfeld and the world oceans amounts to 40–50 Pg per annum Falkowski, 1997a; Longhurst et al., 1995). (Figure 3 and Table 4). Although estimates vary between models, based In contrast to terrestrial ecosystems, the funda- on how the parameters are derived, for illustrative mental limitation of primary production in the purposes we use a model based on empirical ocean is not irradiance per se, but temperature parametrization of the daily integrated photosyn- and the concentration of chlorophyll in the upper thesis profiles as a function of depth. The physical ocean. The latter is a negative feedback; i.e., the depth at which 1% of irradiance incident on the more the chlorophyll in the water column, sea surface remains is called the euphotic zone. the shallower is the euphotic zone. Hence, to This depth can be calculated from surface double NPP requires nearly a fivefold increase in chlorophyll concentrations, and defines the base chlorophyll concentration. of the water column at which net photosynthesis can be supported. Given such information, net primary production can be calculated following 8.05.4.4 Quantum Efficiency of NPP the general equation: The photosynthetically available radiation b 18 PPeu ¼ Csat ·Zeu ·Popt · DL ·F ð6Þ (400–700 nm) for the world oceans is 4.5 £ 10 mol of photons per annum, which is .9.8 £ 1020 21 where PPeu is daily net primary production kJ yr . The average energy stored by photosyn- integrated over the euphotic zone, Csat is the thetic organisms amounts to ,39 kJ per gram of satellite-based (upper water column; i.e., derived carbon fixed (Platt and Irwin, 1973). Given an b from Table 3) chlorophyll concentration, Popt is annual net production of 40 Pg C for phytoplank- the maximum daily photosynthetic rate within the ton, and an estimated production of 4 Pg yr21 by water column, Zeu is the depth of the euphotic benthic photoautotrophs, the photosynthetically zone, DL is the photoperiod (Behrenfeld and stored radiation is equal to ,1.7 £ 1018 kJ yr21. Falkowski, 1997a), and F is a function describing The fraction of photosynthetically available solar the shape of the photosynthesis depth profile. energy conserved by photosynthetic reactions in

Table 4 Annual and seasonal net primary production (NPP) of the major units of the biosphere. Ocean NPP Land NPP

Seasonal April–June 10.9 15.7 July–September 13.0 18.0 October–December 12.3 11.5 January–March 11.3 11.2 Biogeographic Oligotrophic 11.0 Tropical rainforests 17.8 Mesotrophic 27.4 Broadleaf deciduous forests 1.5 Eutrophic 9.1 Broadleaf and needleleaf forests 3.1 Macrophytes 1.0 Needleleaf evergreen forests 3.1 Needleleaf deciduous forest 1.4 Savannas 16.8 Perennial grasslands 2.4 Broadleaf shrubs with bare soil 1.0 Tundra 0.8 Desert 0.5 Cultivation 8.0 Total 48.5 56.4

Source: Field et al. (1998). After Field et al. (1998). All values in GtC. Ocean color data are averages from 1978 to 1983. The land vegetation index is from 1982 to 1990. Ocean NPP estimates are binned into three biogeographic categories on the basis of annual average Csat for each satellite : 23 : 23 23 pixel, such that oligotrophic ¼ Csat , 0 1mgm , mesotrophic ¼ 0 1 , Csat , 1mgm , and eutrophic ¼ Csat . 1mgm (Antoine et al., 1996). This estimate includes a 1 GtC contribution from macroalgae (Smith, 1981). Differences in ocean NPP estimates between Behrenfeld and Falkowski (1997) and those in the global annual NPP for the biosphere and this table result from: (i) addition of Arctic and Antarctic monthly ice masks; (ii) correction of a rounding error in previous calculations of pixel area; and (iii) changes in the designation of the seasons to correspond with Falkowski et al. (1998). 196 Biogeochemistry of Primary Production in the Sea the world oceans amounts to 1.7 £ 1018 kJ/ marine phytoplankton is rapidly consumed by 9.8 £ 1020 ¼ 0.0017 or 0.17%. Thus, on average, grazers or sinks and is transferred from the surface in the oceans, 0.0007 mol C is fixed per mole of ocean to the ocean interior. Upon entering the incident photons; this is equivalent to an effective ocean interior, virtually all of the organic matter quantum requirement of 1,400 quanta per CO2 is oxidized by heterotrophic microbes, and in the fixed. This value is less than 1% of the theoretical process is converted back to inorganic carbon. maximum quantum efficiency of photosynthesis; Elucidating how this transfer occurs, what controls the relatively small realized efficiency is due to the it, how much carbon is transferred via this mecha- fact that photons incident on the ocean surface nism, on what timescales, and whether the process have a small probability of being absorbed by is in steady state was a major focus for research in phytoplankton before they are either absorbed by the latter portion of the twentieth century. water or other molecules (e.g., organic matter), or are scattered back to space. 8.05.5 EXPORT, NEW AND “TRUE NEW” 8.05.4.4.1 Comparing efficiencies for oceanic PRODUCTION and terrestrial primary production We can imagine that NPP produced by photo- The average chlorophyll concentration of autotrophs in the upper, sunlit regions of the ocean the world ocean is 0.24 mg m23 and the (the euphotic zone) is consumed in the same average euphotic zone depth is 54 m; thus the general region by heterotrophs. In such a case, the average integrated chlorophyll concentration is basic reaction given by Equation (4) is simply ,13 mg m22. Carbon to chlorophyll ratios of balanced in the reverse direction due to respiration phytoplankton typically range between 40:1 and by heterotrophic organisms, and no organic matter 100:1 by weight (Banse, 1977). Given the total leaves the ecosystem. This very simple “balanced area of the ocean of 3.1 £ 108 km2, the total state” model, also referred to as the microbial loop carbon biomass in phytoplankton is 0.25–0.65 Pg. (e.g., Azam, 1998), accounts for the fate of most of If NPP is ,40 Pg per annum, and assuming the the organic matter in the oceans (and on time- ocean is in steady state (a condition we will scales of decades, terrestrial ecosystems as well). discuss in more detail), the living phytoplankton In marine ecology, this process is sometimes biomass turns over 60–150 times per year, which called “regenerated production”; i.e., organic is equivalent to a turnover time of 2–6 d. In matter produced by photoautotrophs is locally þ contrast, terrestrial plant biomass amounts to regenerated to inorganic nutrients (CO2,NH4 , 22 ,600–1,000 Pg C, most of which is in the form PO3 ) by heterotrophic respiration. It should be of wood (Woodwell et al., 1978). Estimates of noted here that with the passage of organic matter terrestrial plant NPP are in the range of 50–65 Pg from one level of a marine food chain to the next C per annum, which gives an average turnover (e.g., from primary producer to heterotrophic time of ,12–20 yr (Field et al., 1998). Thus, the consumer), a metabolic “tax” must be paid in the flux of carbon through aquatic photosynthetic form of respiration, such that the net metabolic organisms is about 1,000-fold faster than terres- potential of the heterotrophic biomass is always trial ecosystems, while the storage of carbon in the less than that of the primary producers. This does latter is about 1,000-fold higher than the former. not mean, a priori, that photoautotrophic biomass Moreover, the total photon flux to terrestrial is always greater, as heterotrophs may grow environments amounts to ,2 £ 1018 mol yr21, slowly and accumulate biomass; however, as which gives an effective quantum yield of heterotrophs grow faster, their respiratory rates ,0.002. In other words, on average one CO2 must invariably increase. The rate of production molecule is fixed for every 500 incident photons. (i.e., the energy flux) of heterotrophic biomass is The results of these calculations suggest that always constrained by NPP. terrestrial vegetation is approximately three times Let us imagine a second scenario. Some more efficient in utilizing incident solar radiation fraction of the primary producers and/or hetero- to fix carbon than are aquatic photoautotrophs. trophs sink below a key physical gradient, such as This situation arises primarily from the relative a thermocline, and for whatever reason cannot paucity of aquatic photoautotrophs in the ocean ascend back into the euphotic zone. If the water and the fact that they must compete with the column is very deep, sinking organic matter will media (water) for light. most likely be consumed by heterotrophic This comparison points out a fundamental diffe- microbes in the ocean interior. The flux of organic rence between the two ecosystems in the context of carbon from the euphotic zone is often called the global carbon cycle. On timescales of decades export production, a term coined by Wolfgang to centuries, carbon fixed in terrestrial ecosystems Berger. Export production is an important conduit can be temporarily stored in organic matter (e.g., for the exchange of carbon between the upper forests), whereas most of the carbon fixed by ocean and the ocean interior (Berger et al., 1987). Export, New and “True New” Production 197 This conduit depletes the upper ocean of inorganic maintaining the steady-state levels of atmospheric carbon and other nutrients essential for photosyn- CO2 (Sarmiento et al., 1992; Siegenthaler and thesis and the biosynthesis of organic matter. In Sarmiento, 1993). the central ocean basins, export production is a relatively small fraction of total primary pro- duction, amounting to 5–10% of the total carbon 8.05.5.1 Steady-state versus Transient State fixed per annum (Dugdale and Wilkerson, 1992). At high latitudes and in nutrient-rich areas, The concepts of new, regenerated, and export however, diatoms and other large, heavy cells production are central to understanding many can form massive blooms and sink rapidly. In aspects of the role of aquatic photosynthetic such regions, export production can account organisms in biogeochemical cycles in the oceans. for 50% of the total carbon fixation (Bienfang, In steady state, the globally averaged fluxes of 1992; Campbell and Aarup, 1992; Sancetta et al., new nutrients must match the loss of the nutrients 1991; Walsh, 1983). The subsequent oxidation contained in organic material. If this were not so, and remineralization of the exported produc- there would be a continuous depletion of nutrients tion enriches the ocean interior with inorganic in the euphotic zone and photoautotrophic bio- carbon by ,200 mM in excess of that which mass and primary production would slowly would be supported solely by air–sea exchange decline (Eppley, 1992). Thus, in the steady state, (Figure 4 and Table 5). This enrichment is called the sinking fluxes of organic nitrogen and the the biological pump (Broecker et al., 1980; production of N2 by denitrifying bacteria must Sarmiento and Bender, 1994; Volk and Hoffert, equal the sum of the upward fluxes of inorganic 1985). The biological pump is crucial to nitrogen, nitrogen fixation, and the atmospheric

Figure 4 Vertical profiles of total dissolved inorganic carbon (TIC) in the ocean. Curve A corresponds to a theoretical profile that would have been obtained prior to the Industrial Revolution with an atmospheric CO2 21 concentration of 280 mmol mol . The curve is derived from the solubility coefficients for CO2 in seawater, using a typical thermal and salinity profile from the central Pacific Ocean, and assumes that when surface water cools and sinks to become deep water it has equilibrated with atmospheric CO2. Curve B corresponds to the same calculated 21 solubility profile of TIC, but in the year 1995, with an atmospheric CO2 concentration of 360 mmol mol . The difference between these two curves is the integrated oceanic uptake of CO2 from anthropogenic emissions since the beginning of the Industrial Revolution, with the assumption that biological processes have been in steady state (and hence have not materially affected the net influx of CO2). Curve C is a representative profile of measured TIC from the central Pacific Ocean. The difference between curve C and B is the contribution of biological processes to the uptake of CO2 in the steady state (i.e. the contribution of the “biological pump” to the TIC pool.) (courtesy of Doug Wallace and the World Ocean Circulation Experiment). 198 Biogeochemistry of Primary Production in the Sea Table 5 Export production and ef ratios calculated (i.e., oxidation) of the elements in organic matter from the model of Laws et al. (2000). produced in the open ocean. Further, as carbon and nitrogen in living organisms are largely found Export ef 21 in chemically reduced forms, while remineralized (GtC y ) forms are virtually all oxidized, the remineraliza- Ocean basin tion of organic matter was coupled to the depletion Pacific 4.3 0.19 of oxygen. The relationship could be expressed Atlantic 4.3 0.25 stoichiometrically as Indian 1.5 0.15 22 Antarctic 0.62 0.28 ½106ðCH2OÞ16NH31PO4 þ138O2 Arctic 0.15 0.56 ! 106CO þ 122H O þ 16NO2 Mediterranean 0.19 0.24 2 2 3 22 þ Global 11.1 0.21 þ PO4 þ 16H ð7Þ Total production Hidden within this balanced chemical formulation Oligotrophic 1.04 0.15 are biochemical redox reactions, which are (chl a , 0.1 mg m23) Mesotrophic 6.5 0.18 contained within specific groups of organisms. (0.1 # chl a , 1.0 mg m23) (In the oxidation of organic matter, there is some Eutrophic 3.6 0.36 ambiguity about the stoichiometry of O2/P. (chl a $ 1.0 mg m23) Assuming that the mean oxidation level of organic carbon is that of carbohydrate (as is the case in Ocean depth Equation (9)), then the oxidation of that carbon is 0 2 100 m 2.2 0.31 100 m–1 km 1.4 0.33 equimolar with O. Alternatively, some organic .1 km 7.4 0.18 matter may be more or less reduced than carbohydrate, and therefore require more or less O for oxidation. Note also that the oxidation of 2 NH3 to NO3 requires four atoms of O, and leads to þ deposition of fixed nitrogen in the form of aero- the formation of one H2O and one H .) When the sols (the latter is produced largely as a conse- reactions primarily occur at depth, Equation (7) is quence of air pollution and, to a lesser extent, from driven to the right, while when the reactions lightning). primarily occur in the euphotic zone, they are driven to the left. Note that in addition to reducing CO2 to organic matter, formation of organic matter by photoautotrophs requires reduction of 8.05.6 NUTRIENT FLUXES nitrate to the equivalent of ammonia. These two forms of nitrogen are critically important in Primary producers are not simply sacks of helping to quantify “new” and export production. organic carbon. They are composed of six major elements, namely, hydrogen, carbon, oxygen, nitrogen, phosphorus, and sulfur, and at least 54 other trace elements and metals (Schlesinger, 8.05.7 NITRIFICATION 1997). In steady state, the export flux of organic 2 matter to the ocean interior must be coupled to the NO3 is produced via oxidation of NH3 by a specific group of eubacteria, the nitrifiers, that are upward flux of several of these essential nutrients. oblibate aerobes found primarily in the water The fluxes of nutrients are related to the elemental stoichiometry of the organic matter that sinks column. The oxidation of NH3 is coupled to the reduction of inorganic carbon to organic matter; into the ocean interior. This relationship, first hence nitrification is an example of a chemoauto- pointed out by Alfred Redfield in 1934, was based on the chemistry of four of the major elements trophic process that couples the aerobic nitrogen in the ocean, namely, carbon, nitrogen, phos- cycle to the carbon cycle. However, because the thermodynamic gradient is very small, the effi- phorus, and oxygen. ciency of carbon fixation by nitrifying bacteria is low and does not provide an ecologically 8.05.6.1 The Redfield Ratio significant source of organic matter in the oceans. In the contemporary ocean, global CO2 fixation by In the ocean interior, the ratio of fixed inorganic marine nitrifying bacteria only amounts to nitrogen (in the form of NO32)toPO4 in the ,0.2 Pg C per annum, or ,0.5% of marine dissolved phase is remarkably close to the ratios of photoautotrophic carbon fixation. the two elements in living plankton. Hence, it There are two major sources of nutrients in the seemed reasonable to assume that the ratio of the euphotic zone. One is the local regeneration of þ two elements in the dissolved phase was the result simple forms of combined elements (e.g., NH4 , 22 22 of the sinking and subsequent remineralization HPO4 ,SO4 ) resulting from the metabolic Nitrification 199 activity of metazoan and microbial degradation. relative to the source carbon isotopic value. The The second is the influx of distantly produced, extent of the actual fractionation is somewhat “new” nutrients, imported from the deep ocean, variable and is a function of carbon availability the atmosphere (i.e., nitrogen fixation, atmos- and of the transport processes for inorganic carbon pheric pollution), or terrestrial runoff from into the cells, as well as the specific carboxylation streams, rivers, and estuaries (Dugdale and pathway (Laws et al., 1997). However, regardless Goering, 1967). In the open ocean, these two of the quantitative aspects, the net effect of carbon sources can be usefully related to the form of fixation is an enrichment of the inorganic carbon inorganic nitrogen assimilated by phytoplankton. pool in 13C, while the organic carbon produced is Because biological nitrogen fixation is relatively enriched in 12C. low in the ocean (see below) and nitrification in the upper mixed layer is sluggish relative to the 8.05.7.2 Carbon Isotope Fractionation in Organic assimilation of nitrogen by photoautotrophs, Matter and Carbonates nitrogen supplied from local regeneration is assimilated before it has a chance to become The isotopic fractionation in carbonates mirrors oxidized. Hence, regenerated nitrogen is primarily the relative amount of organic carbon buried. It is in the form of ammonium or urea. In contrast, the generally assumed that the source carbon, from fixed inorganic nitrogen in the deep ocean has vulcanism (the so-called “mantle” carbon) has an sufficient time (hundreds of years) to become isotopic value of approximately 25‰. As mass oxidized, and hence the major source of new balance must constrain the isotopic signatures of 15 nitrogen is in the form of nitrate. Using NH4 and carbonate carbon and organic carbon with the 15 NO3 as tracers, it is possible to estimate the mantle carbon, then fraction of new nitrogen that fuels phytoplankton production (Dugdale and Wilkerson, 1992). This dw 2 dcarb forg ¼ ð8Þ approach provides an estimate of both the up- DB 15 2 ward flux of nitrate required to sustain the NO3 supported production, as well as the downward where forg is the fraction of organic carbon buried, flux of organic carbon, which is required to dw is the average isotopic content of the carbon maintain a steady-state balance (Dugdale and weathered, dcarb is the isotopic signature of the Wilkerson, 1992; Eppley and Peterson, 1979). carbonate carbon, and DB the isotopic difference between organic carbon and carbonate carbon deposited in the ocean. Equation (8) is a steady- 8.05.7.1 Carbon Burial state model that presumes the source of carbon from the mantel is constant over geological time. On geological timescales, there is one import- This basic model is the basis of nearly all ant fate for NPP, namely, burial in the sediments. estimates of organic carbon burial rates (Berner By far the largest reservoir of organic matter on et al., 1983; Kump and Arthur, 1999). Earth is locked up in rocks (Table 1). Virtually all Carbonate isotopic analyses reveal positive of this organic carbon is the result of the burial of excursions (i.e., implying organic carbon burial) exported marine organic matter in coastal sedi- in the Proterozioc, and more modest excursions ments over literally billions of years of Earth’s throughout the Phanaerozic (Figure 5). Burial of history. On geological timescales, the burial of organic carbon on geological timescales implies marine NPP effectively removes carbon from that export production must deviate from the biological cycles, and places most (not all) of that steady state on ecological timescales. Such a carbon into the slow carbon cycle. A small deviation requires changing one or more of fraction of the organic matter escapes tectonic (i) ocean nutrient inventories, (ii) the utilization processing via the Wilson cycle and is perma- of unused nutrients in enriched areas, (iii) the nently buried, mostly in continental rocks. The average elemental composition of the organic burial of organic carbon is inferred not from direct material, or (iv) the “rain” ratios of particulate measurement, but rather from indirect means. One organic carbon to particulate inorganic carbon to of the most common proxies used to derive burial the seafloor. on geological timescales is based on isotopic fractionation of carbonates. The rationale for this analysis is that the primary enzyme responsible 8.05.7.3 Balance between Net Primary for inorganic carbon fixation is ribulose 1,5- Production and Losses bisphosphate carboxylase/oxygenase (RuBisCO), which catalyzes the reaction between ribulose 1,5- In the ecological theater of aquatic ecosystems, 2 bisphosphate and CO2 (not HCO3 ), to form two the observed photoautotrophic biomass at any molecules of 3-phosphoglycerate. The enzyme moment in time represents a balance between the strongly discriminates against 13C, such that the rate of growth and the rate of removal of that resulting isotopic fractionation amounts to ,27‰ trophic level. The burial of organic carbon in 13 Figure 5 Phanerozoic d Ccarb record. The Jurassic through the Cenozoic record was generated from bulk sediment carbonates primarily from open ocean Atlantic Deep Sea Drilling Project boreholes; Lower Jurassic samples were used from the Mochras Borehole (Wales) (Katz et al., in review). Dashed intervals indicate data gaps. Singular spectrum analysis was used to generate the Mesozoic–Cenozoic curve. The Paleozoic record was generated from brachiopods (Veizer et al., 1999; note that the timescale has been adjusted from the original reference by interpolating between period boundaries). Data were averaged for each time slice to obtain the Paleozoic curve. The timescales of Berggren et al. (1995; Cenozoic), Gradstein et al. (1995; Mesozoic), and GSA (Paleozoic) were used. Limiting Macronutrients 201 the lithosphere requires that the ecological 8.05.8 LIMITING MACRONUTRIENTS balance between NPP and respiration diverge; i.e., the global ocean must be net autotropic. For A general feature of aquatic environments is simplicity, we can express the time-dependent that because the oxidation of organic nutrients to change in photoautrophic biomass by a linear their inorganic forms occurs below the euphotic differential equation: zone where the competing processes of assimila- tion of nutrients by photoautotrophs do not occur, dP=dt ¼½Pðm 2 mÞð9Þ the pools of inorganic nutrients are much higher at depth. As the only natural source of photosynthe- where ½P is photoautrophic biomass (e.g., organic tically active radiation is the sun, the gradients of carbon), m is the specific growth rate (units of light and nutrients are from opposite directions. 1/time), and m is the specific mortality rate (units Thermal or salinity differences in the surface of 1/time). In this equation, we have lumped all layers produce vertical gradients in density that mortality terms, such as grazing and sinking effectively retard the vertical fluxes of soluble together into one term, although each of these nutrients from depth. Thus, in the surface layers of loss processes can be given explicitly (Banse, a stratified water column, nutrients become 1994). Two things should be noted regarding depleted as the photoautotrophs consume them Equation (9). First, m and m are independent at rates exceeding their rate of vertical supply. variables; i.e., changes in P can be independently Indeed, throughout most of the world oceans, the ascribed to one or the other process. Second, by concentrations of dissolved inorganic nutrients, definition, a steady state exists when dP=dt is zero. especially fixed inorganic nitrogen and phosphate, are exceedingly low, often only a few nM. One or the other of these nutrients can limit primary 8.05.7.4 Carbon Burial in the Contemporary production. However, the concept of limitation Ocean requires some discussion. The burial of organic carbon in the modern oceans is primarily confined to a few regions where the supply of sediments from terrestrial 8.05.8.1 The Two Concepts of Limitation sources is extremely high. Such regions include the Amazon outfall and Indonesian mud belts. In The original notion of limitation in ecology was contrast, the oxidation of organic matter in the related to the yield of a crop. A limiting factor was interior of the contemporary ocean is extremely the substrate least available relative to the efficient; virtually no carbon is buried in the deep requirement for synthesis of the crop (Liebig, sea. Similarly, on most continental margins, 1840). This concept formed a strong underpinning organic carbon that reaches the sediments is of agricultural chemistry and was used to design consumed by microbes within the sediments, the elemental composition of fertilizers for such that very little is actually buried in the commercial crops. This concept subsequently contemporary ocean (Aller, 1998). The solution to was embraced by ecologists and geochemists as Equation (9) must be close to zero; and conse- a general “law” (Odum, 1971). quently, in the absence of human activities, the Nutrients can also limit the rate of growth of oxygen content of Earth’s atmosphere is very photoautotrophs (Blackman, 1905; Dugdale, close to steady state and has been zso for tens, if 1967). Recall that if organisms are in balanced not hundreds, of millions of years. growth, the rate of uptake of an inorganic nutrient relative to the cellular concentration of the nutrient defines the growth rate (Herbert et al., 1956). The uptake of inorganic nutrients is a hyperbolic 8.05.7.5 Carbon Burial in the Precambrian function of the nutrient concentration and can be Ocean conveniently described by a hyperbolic expression In contrast, carbon burial in the Proterozoic of the general form ocean must have occurred as oxygen increased in ; = ; the atmosphere; i.e., the global solution to V ¼ðVmaxðU …ÞÞ ðKs þðU …ÞÞ ð10Þ Equation (9) must have been .0. Photoauto- trophic biomass could have increased until some where V is the instantaneous rate of nutrient ; element became limiting. Thus, the original uptake, Vmax is the maximum uptake rate, (U …) feedback between the production of photoauto- represent the substrate concentration of nutrient U, trophic biomass in the oceans and the atmospheric etc., and Ks is the concentration supporting half content of oxygen was determined by an element the maximum rate of uptake (Dugdale, 1967; that limited the crop size of the photoautotrophs in Monod, 1942). There can be considerable the Archean or Proterozoic ocean. What was that variation between species with regard to Ks and element, and why did it become limiting? Vmax values and these variations are potential 202 Biogeochemistry of Primary Production in the Sea sources of competitive selection (Eppley et al., aluminum and transition metals such as iron are 1969; Tilman, 1982). mediated at either low salinity, low pH, or high It should be noted that Liebig’s notion of oxidation states of the cations (Stumm and limitation was not related a priori to the intrinsic Morgan, 1981). Although these reactions would rate of photosynthesis or growth. For example, reduce the overall soluble phosphate concen- photosynthetic rates can be (and often are) limited tration, the initial condition of the Archean by light or temperature. The two concepts of ocean probably had a fixed low N:P ratio in the limitation (yield and rate) are often not understood dissolved inorganic phase. As N2 fixation pro- correctly: the former is more relevant to biogeo- ceeded, that ratio would have increased with a chemical cycles, the latter is more critical to buildup of ammonium in the ocean interior. The selection of species in ecosystems. accumulation of fixed nitrogen in the oceans would continue until the N:P ratio of the inorganic elements reached equilibrium with the N:P ratio of 8.05.9 THE EVOLUTION OF THE NITROGEN the sedimenting particulate organic matter (POM). CYCLE Presumably, the latter ratio would approximate that of extant, nitrogen-fixing marine cyanobac- Globally, nitrogen and phosphorus are the two teria, which is ,16:1 by atoms (Copin-Montegut elements that immediately limit, in a Liebig sense, and Copin-Montegut, 1983; Redfield, 1958; the biologically mediated carbon assimilation in Quigg et al., 2003) or greater (Letelier et al., the oceans by photoautotrophs. It is frequently 1996) and would ultimately be constrained by argued that since N2 is abundant in both the ocean the availability of phosphate (Falkowski, 1997; and the atmosphere, and, in principle, can be Tyrell, 1999). biologically reduced to the equivalent of NH3 by The formation of nitrate from ammonium by N2-fixing cyanobacteria, nitrogen cannot be limit- nitrifying bacteria requires molecular oxygen; ing on geological timescales (Barber, 1992; hence, nitrification must have evolved following Broecker et al., 1980; Redfield, 1958). Therefore, the formation of free molecular oxygen in the phosphorus, which is supplied to the ocean by the oceans by oxygenic photoautotrophs. Therefore, weathering of continental rocks, must ultimately from a geological perspective, the conversion of limit biological productivity. The underlying ammonium to nitrate probably proceeded rapidly assumptions of these tenets should, however, be 2 and provided a substrate, NO3 , that eventually considered within the context of the evolution of could serve both as a source of nitrogen for biogeochemical cycles. photoautotrophs and as an electron acceptor for a By far, the major source of fixed inorganic diverse group of heterotrophic, anaerobic bacteria, nitrogen for the oceans is via biological nitrogen the denitrifiers. fixation. Although in the Archean atmosphere, In the sequence of the three major biological electrical discharge or bolide impacts may have processes that constitute the nitrogen cycle, promoted NO formation from the reaction denitrification must have been the last to emerge. between N and CO , the yield for these reactions 2 2 This process, which permits the reduction of NO2 is extremely low. Moreover, atmospheric NH 3 3 to (ultimately) N , occurs in the modern ocean in would have photodissociated from UV radiation 2 three major regions, namely, continental margin (Kasting, 1990), while N would have been stable 2 sediments, areas of restricted circulation such as (Kasting, 1990; Warneck, 1988). Biological N2 fixation is a strictly anaerobic process (Postgate, fjords, and oxygen minima zones of perennially 1971), and the sequence of the genes encoding the stratified seas (Christensen et al., 1987; Codispoti catalytic subunits for nitrogenase is highly con- and Christensen, 1985; Devol, 1991; Nixon et al., served in cyanobacteria and other eubacteria, 1996). In all cases, the process requires hypoxic or strongly suggesting a common ancestral origin anoxic environments and is sustained by high (Zehr et al., 1995). The antiquity and homology of sinking fluxes of organic matter. Denitrification nitrogen fixation capacity also imply that fixed appears to have evolved independently several inorganic nitrogen was scarce prior to the times; the organisms and enzymes responsible for evolution of diazotrophic organisms; i.e., there the pathway are highly diverse from a phylogenetic was strong evolutionary selection for nitrogen and evolutionary standpoint. fixationintheArcheanorearlyProterozoic With the emergence of denitrification, the ratio periods. In the contemporary ocean, N2 is still of fixed inorganic nitrogen to dissolved inorganic catalyzed solely by prokaryotes, primarily cyano- phosphate in the ocean interior could only be bacteria (Capone and Carpenter, 1982). depleted in nitrogen relative to the sinking flux of While apatite and other calcium-based and the two elements in POM. Indeed, in all of the substituted solid phases of phosphate minerals major basins in the contemporary ocean, the N:P precipitated in the primary formation of crustal ratio of the dissolved inorganic nutrients in the sediments, secondary reactions of phosphate with ocean interior is conservatively estimated at 14.7 Functional Groups 203

2 32 Figure 6 (a)–(c) Vertical profiles of NO3 /HPO4 ratios in each of the three major ocean basins. The data were taken from the GEOSECS database. In all three basins, the N:P ratio converges on an average value that is significantly lower than the 16:1 ratio predicted by Redfield. The deficit in nitrogen relative to phosphorus is 2 32 presumed to be a result of denitrification. Note that in the upper 500 m of the water column, NO3 /HPO4 ratios generally decline except in a portion of the Atlantic that corresponds to the eastern Mediterranean.

by atoms (Fanning, 1992) or less (Anderson and phytoplankton (Lipps, 1993). The Redfield N:P Sarmiento, 1994)(Figure 6). ratio of 16:1 for POM (Codispoti, 1995; Copin- There are three major conclusions that may be Montegut and Copin-Montegut, 1983; McElroy, drawn from the foregoing discussion: 1983; Redfield et al., 1963, 1958) is an upper (i) Because the ratio of the sinking flux of bound, which is not observed for the two elements particulate organic nitrogen and particulate phos- in the dissolved inorganic phase in the ocean phorus exceeds the N:P ratio of the dissolved pool interior. The deficit in dissolved inorganic fixed of inorganic nutrients in the ocean interior, the nitrogen relative to soluble phosphate in the ocean average upward flux of inorganic nutrients must be represents a slight imbalance between nitrogen slightly enriched in phosphorus relative to nitrogen fixation and denitrification on timescales of as well as to the elemental requirements of the ,103–104 yr (Codispoti, 1995). photoautotrophs (Gruber and Sarmiento, 1997; (iii) If dissolved inorganic nitrogen rather than Redfield, 1958). Hence, although there are some phosphate limits productivity in the oceans, then it exceptions (Kromer, 1995; Wu et al., 2000), follows that the ratio of nitrogen fixation/denitri- dissolved, inorganic fixed nitrogen generally limits fication plays a critical role in determining the net primary production throughout most of the world’s biologically mediated exchange of CO2 between oceans (Barber, 1992; Falkowski et al., 1998). the atmosphere and ocean (Codispoti, 1995). (ii) The N:P ratio of the dissolved pool of inorganic nutrients in the ocean interior was established by biological processes, not vice 8.05.10 FUNCTIONAL GROUPS versa (Redfield et al., 1963, 1934). The elemental composition of marine photoautotrophs has been As we have implied throughout the foregoing conserved since the evolution of the eukaryotic discussion, the biologically mediated fluxes of 204 Biogeochemistry of Primary Production in the Sea elements between the upper ocean and the ocean by silicious organisms, however, the ocean is interior are critically dependent upon key groups relatively depleted in dissolved silica. Although of organisms. Fluxes between the atmosphere and diatom frustules (their silicified cell walls) tend to ocean, as well as between the ocean and the dissolve and are relatively poorly preserved in lithosphere, are mediated by organisms that marine sediments, enough silica is buried to keep catalyze phase state transitions from either gas to the seawater undersaturated, throughout the ocean. solute/solid or from solute to solid/gas phases. For As the residence time of silica in the oceans is example, autotrophic carbon fixation converts ,104 yr (i.e., about an order of magnitude longer gaseous CO2 to a wide variety of organic carbon than the mean deep-water circulation), one can get molecules, virtually all of which are solid or an appreciation for the silicate demands and dissolved solids at physiological temperatures. regeneration rates by following the concentration Respiration accomplishes the reverse. Nitrogen gradients of dissolved silica along isopycnals. fixation converts gaseous N2 to ammonium and While these demands are generally attributed to thence to organic molecules, while denitrification diatoms, radiolarians (a group of nonphotosyn- accomplishes the reverse. Calcification converts thetic, heterotrophic protists with silicious tests dissolved inorganic carbon and calcium to solid that are totally unrelated to diatoms) are not phase calcite and aragonite, whereas silicification uncommon, and radiolarian shells are abundant in converts soluble silicic acid to solid hydrated the sediments of Southern Ocean. Silica is also amorphous opal. Each of these biologically precipitated by various sponges and other protists. catalyzed processes is dependent upon specific As a functional group, the silicate precipitators are metabolic sequences (i.e., gene families encoding identified by their geochemical signatures in the a suite of enzymes) that evolved over hundreds of sediments and in the silica chemistry of the oceans. millions of years of Earth’s history, and have, over corresponding periods, led to the massive accumu- lation of oxygen in the atmosphere, and opal, 8.05.10.2 Calcium Carbonate Precipitation carbonates, and organic matter in the lithosphere. Like silica precipitation, calcium carbonate is Presumably, because of parallel evolution as well not confined to a specific phylogenetically distinct as lateral gene transfer, these metabolic sequences group of organisms, but evolved (apparently have frequently co-evolved in several groups of independently) several times in marine organisms. organisms that, more often than not, are not Carbonate sediments blanket much of the Atlantic closely related from a phylogenetic standpoint basin, and are formed from the shells of both (Falkowski, 1997). Based on their biogeochemical coccolithophorids and foraminifera (Milliman, metabolism, these homologous sets of organisms 1993). (In the Pacific, the carbon compensation are called functional groups or biogeochemical depth is generally higher than the bottom, and guilds; i.e., organisms that are related through hence, in that basin carbonates tend to dissolve common biogeochemical processes rather than a rather than become buried.) As the crystal common evolutionary ancestor affiliation. structure of the carbonates in both groups is calcite (as opposed to the more diagenetically 8.05.10.1 Siliceous Organisms susceptible aragonite), the preservation of these minerals and their co-precipitating trace elements In the contemporary ocean, the export of provides an invaluable record of ocean history. particulate organic carbon from the euphotic Although on geological timescales huge amounts zone is highly correlated with the flux of of carbon are removed from the atmosphere and particulate silicate. Most of the silicate flux is a ocean and stored in the lithosphere as carbonates, consequence of precipitation of dissolved ortho- on ecological timescales, carbonate formation silicic acid by diatoms to form amorphous opal leads to the formation of CO2. This reaction can that makes up the cell walls of these organisms. be summarized by the following: These hard-shelled cell walls presumably help the 2 2þ organisms avoid predation, or if ingested, increase 2HCO3 þ Ca ! CaCO3 þ CO2 þ H2O the likelihood of intact gut passage through some ð11Þ metazoans (Smetacek, 1999). In precipitating silicate, diatoms simultaneously fix carbon. Unlike silicate precipitation, calcium carbonate Upon depleting the euphotic zone of nutrients, precipitation leads to strong optical signatures that the organisms frequently sink en masse, and while can be detected both in situ and remotely (Balch some are grazed en route, many sink as intact et al., 1991; Holligan and Balch, 1991). The basic cells. Ultimately, either fate leads to the gravi- principle of detection is the large, broadband (i.e., tationally driven export flux of particulate organic “white”) scattering cross-sections of calcite. The carbon into the ocean interior. high scattering cross-sections are detected by Silica is supplied to the oceans from the weath- satellites observing the upper ocean as relatively ering of continental rocks. Because of precipitation highly reflective properties (i.e., a “bright” ocean). Functional Groups 205 Using this detection scheme, one can reconstruct The geological record during the Pleistocene global maps of planktonic calcium carbonate reveals a periodicity of opal/calcite deposition precipitating organisms in the upper ocean. corresponding to glacial/interglacial periods. Such In situ analysis can be accompanied by optical alterations in mineral deposition are probably rotation properties (polarization) to discriminate related to upper ocean turbulence; i.e., the calcite from other scattering particles. In situ pro- sedimentary record is a “fax” machine of mixing files of calcite can be used to construct the vertical (Falkowski, 2002). Glacial periods appear to be distribution of calcium carbonate-precipitating characterized by higher wind speeds and a planktonic organisms that would otherwise not stronger thermal contrast between the equator be detected by satellite remote sensing because and the poles. These two factors would, in they are too deep in the water column. accordance with the simple nutrient uptake Over geological time, the relative abundances model, favor diatoms over coccolithophores. of key functional groups change. For example, During interglacials, more intense ocean stratifi- relative coccolithophorid abundances generally cation, weaker winds, and a smaller thermal increased through the Mesozoic, and underwent a contrast between the equator and the poles culling at the Cretaceous Tertiary (K/T) boundary, would tend to reduce upper-ocean mixing and followed by a general waning throughout the favor coccolithophores (Iglesias-Rodriguez et al., Cenozoic. The changes in the coccolithophorid 2002). While other factors such as silica avail- abundances appear to follow eustatic sea-level ability undoubtedly also influenced the relative variations, suggesting that transgressions lead to success of diatoms and coccolithophores on these higher calcium carbonate deposition. In contrast, timescales, we suggest that the climatically forced diatom sedimentation increases with regressions cycle, played out on timescales of 40 kyr and and, since the K/T impact, diatoms have generally 100 kyr (over the past 1.9 Myr), can be understood replaced coccolithophorids as ecologically import- as a long-term competition that never reaches ant eukaryotic phytoplankton. On much finer an exclusion equilibrium condition (Falkowski timescales during the Pleistocene, it would appear et al., 1998). that interglacial periods favor coccolithophorid Can the turbulence argument be extended to abundance, while glacial periods favor diatoms. even longer timescales to account for the switch in The factors that lead to glacial–interglacial the dominance from coccolithophorids to diatoms variations between these two functional groups in the Cenozoic? The fossil record of diatoms in are relevant to elucidating their distributions in the the Mesozoic is obscured by problems of preser- contemporary ecological setting of the ocean vation; however several species are preserved in (Tozzi, 2001). the late Jurassic (Harwood and Nikolaev, 1995), suggesting that the origins were in the early Jurassic or perhaps as early as the Triassic. It is clear, however, that this group did not contribute 8.05.10.3 Vacuoles nearly as much to export production during In addition to a silicic acid requirement, Mesozoic times. We suggest that the ongoing diatoms, in contrast to dinoflagellates and cocco- successional displacement of coccolithophores by lithophores, have evolved a nutrient storage diatoms in the Cenozoic is, to first order, driven by vacuole (Raven, 1987). The vacuole, which tectonics (i.e., the Wilson cycle). The Mesozoic occupies ,35% of the volume of the cell, can period was relatively warm and was characterized retain high concentrations of nitrate and phos- by a two-cell Hadley circulation, with obliquity phate. Importantly, ammonium cannot be (or is greater than 378, resulting in a thoroughly mixed not) stored in a vacuole. The vacuole allows atmosphere with nearly uniform temperatures diatoms to access and hoard pulses of inorganic over the surface of the Earth. The atmospheric nutrients, thereby depriving potentially competing meridional heat transport decreased the latitudinal groups of these essential resources. Consequently, thermogradients; global winds and ocean circula- diatoms thrive best under eutrophic conditions and tion were both sluggish (Huber et al., 1995). This in turbulent regions where nutrients are supplied relatively quiescent period of Earth’s history was with high pulse frequencies. ideal for coccolithophorids. Following the K/T The competition between diatoms and cocco- impact, and more critically, the onset of polar ice lithophorids can be easily modeled by a resource caps about 32 Ma, the Hadley circulation changed acquisition model based on nutrient uptake dramatically. Presently, there are six Hadley cells, (Equation (9)). In such a model, diatoms domi- and the atmosphere has become drier. The net nate under highly turbulent conditions, when result is more intense thermohaline circulation, their nutrient storage capacity is maximally greater wind mixing and decreased stability advantageous, while coccolithophorids dominate (Barron et al., 1995; Chandler et al., 1992). under relatively quiescent conditions (Tozzi Associated with this decreased stability is the rise et al., 2003). of the diatoms. 206 Biogeochemistry of Primary Production in the Sea Over the past 50 Myr, both carbon and oxygen increase in the efficiency of carbon burial, played isotopic records in fossil foraminifera suggest that a key role in decreasing atmospheric CO2 over the there has been a long-term depletion of CO2 in the past 32 Myr. That biological selection may ocean–atmosphere system and a decrease in influence climate is clearly controversial; howe- temperature in the ocean interior. The result has ver, the trends in succession between taxa on been increased stratification of the ocean, which timescales of tens of millions of years, and cycles has, in turn, led to an increased importance for in dominance on shorter geological timescales beg wind-driven upwelling and mesoscale eddy tur- for explanation. bulence in providing nutrients to the euphotic zone. The ecological dominance of diatoms under sporadic mixing conditions suggests that their long-term success in the Cenozoic reflects an 8.05.11 HIGH-NUTRIENT, increase in event-scale turbulent energy dissipa- LOW-CHLOROPHYLL tion in the upper ocean. But, was the Wilson cycle REGIONS—IRON LIMITATION the only driver? Although weathering of siliceous minerals by On ecological timescales, the biologically CO2 (the so-called “Urey reactions”; but see mediated net exchange of CO2 between the (Berner and Maasch, 1996)) contributed to the ocean and atmosphere is limited by nutrient long-term flux of silica to the oceans (Berner, supply and the efficiency of nutrient utilization 1990), and potentially fostered the radiation of in the euphotic zone. There are three major areas diatoms in the Cenozoic, by itself, orogeny cannot of the world ocean where inorganic nitrogen and explain the relatively sharp increase in diatoms at phosphate are in excess throughout the year, yet the Eocene/Oligocene boundary. Indeed, the the mixed-layer depth appears to be shallower seawater strontium isotope record does not than the critical depth; these are the eastern correspond with these radiations in diatoms equatorial Pacific, the subarctic Pacific, and (Raymo and Ruddiman, 1992). We must look Southern (i.e., Antarctic) Oceans. In the subarctic for other contributing processes. North Pacific, it has been suggested that there is a Shortly after, or perhaps coincident with tight coupling between phytoplankton production Paleocene thermal maximum (55 Ma), was a rise and consumption by zooplankton (Miller et al., in true grasses (Retallack, 2001). This group, 1991). This grazer-limited hypothesis has been which rapidly radiated in the Eocene, rose to used to explain why the phytoplankton in the prominence in Oligocene, a period coincident North Pacific do not form massive blooms in the with a global climatic drying. During this period, spring and summer like their counterparts in however, there was a rapid co-evolution of the North Atlantic (Banse, 1992). In the mid- grazing ungulates that displaced browsers (Janis 1980s, however, it became increasingly clear that and Damuth, 1990). Grasses contain up to 10% the concentration of trace metals, especially iron, dry weight of silica, which forms micromineral was extremely low in all three of these regions deposits in the cell walls; phytoliths (Conley, (Martin, 1991). Indeed, in the eastern equatorial 2002). Indeed, the selection of hypsodont (high Pacific, for example, the concentration of soluble crown) dentition in ungulates from the brachydont iron in the euphotic zone is only 100–200 pM. (leaf eating) early-appearing browsing mammals, Although iron is the most abundant transition coincides with the widespread distribution of metal in the Earth’s crust, in its most commonly phytoliths and grit in grassland forage. It is occurring form, Fe3þ, it is virtually insoluble in tempting to suggest that the rise of grazing seawater. The major source of iron to the euphotic ungulates, which spurred the radiation of grasses, zone is Aeolian dust, originating from continental was, in effect, a biologically catalyzed silicate deserts. In the three major areas of the world weathering process. The deep-root structure of oceans with high inorganic nitrogen in the surface Eocene grasses certainly facilitated silicate mobil- waters and low chlorophyll concentrations, the ization into rivers and groundwaters (Conley, flux of Aeolian iron is extremely low (Duce and 2002). Additionally, upon their annual death and Tindale, 1991). In experiments in which iron was decay, the phytoliths of many temperate grasses artificially added on a relatively large scale to the are potentially transported to the oceans via wind. waters in the equatorial Pacific, Southern Ocean, The feedback between the co-evolution of and subarctic Pacific, there were rapid and mammals and grasses and the supply of silicates dramatic increases in photosynthetic energy con- to the ocean potentially explains the rapid version efficiency and phytoplankton chlorophyll radiation of diatoms, and their continued dom- (Abraham et al., 2000; Behrenfeld et al., 1996; inance in the Cenozoic. There is another potential Kolber et al., 1994; Tsuda et al., 2003). Beyond feedback at play, however, which “locked in” doubt, NPP and export production in all three the diatom preeminence. It is likely that the regions are limited by the availability of a single increase in diatom dominance, and the associated micronutrient—iron. Linking Iron to N2 Fixation 207 8.05.12 GLACIAL–INTERGLACIAL chlorophyll Southern Ocean stimulated phyto- CHANGES IN THE BIOLOGICAL plankton photosynthesis and led to a drawdown of CO2 PUMP atmospheric CO2. Model calculations suggest that the magnitude of this drawdown could have been In the modern (i.e., interglacial) ocean, two cumulatively significant, and accounted for the major factors affect iron fluxes. First, changes in observed variations in atmospheric CO2 recorded land-use patterns and climate over the past several in gases trapped in the ice cores. However, the thousand years have had, and continue to have, sedimentary records reveal large glacial fluxes of marked effects on the areal distribution and extent organic carbon in low- and mid-latitude regions; of deserts. At the height of the Roman Empire areas that are presumably nutrient impoverished. some 2,000 years ago, vast areas of North Africa Was there another factor besides iron addition were forested, whereas today these same areas are to high-nutrient, low-chlorophyll regions that desert. These changes were climatologically contributed to a net export of carbon during induced. Similarly, the Gobi Desert in North glacial times? Central Asia has increased markedly in modern Given that N:P ratios in the ocean interior are times. The flux of Aeolian iron from the Sahara lower than for the sinking flux, an increase in Desert fuels photosynthesis for most of the North the net delivery of fixed inorganic nitrogen to the Atlantic Ocean; that from the Gobi is deposited ocean would also potentially contribute to a net over much of the North Pacific (Duce and Tindale, drawdown of CO . The Antarctic ice core records 1991). The primary source of iron for the Southern 2 suggest that atmospheric CO2 declined from Ocean is Australia, but the prevailing wind vectors ,290 mmol mol21 to ,190 mmol mol21 over a constrain the delivery of the terrestrial dust to period of ,8 £ 104 yr between the interglacial the Indian Ocean. Consequently, the Southern and glacial maxima (Sigman and Boyle, 2000). Ocean is iron limited in the modern epoch Assuming a C:N ratio of about 6.5 by atoms for (Martin, 1990). the synthesis of new organic matter in the euphotic The second factor in this climatological feed- zone, a simple equilibrium, three-box model back is that the major wind vectors are driven by calculation suggests that 600 Pg of inorganic atmosphere–ocean heat gradients. Changes in carbon should have been fixed by marine photo- thermal gradients between the equator and poles autotrophs to account for the change in atmos- lead to changes in wind speed and direction. Wind pheric CO2. This amount of carbon is vectors prior to glaciations appear to have approximately threefold greater than that released supported high fluxes of iron to the Southern to the atmosphere from the cumulative combus- Ocean, thereby presumably stimulating phyto- tion of fossil fuels since the beginning of the plankton production, the export of carbon to Industrial Revolution. The calculated change in depth, and the drawdown of atmospheric CO2 atmospheric CO2 would have required an addition appears to have accompanied glaciations in the of ,1.5 Tg fixed nitrogen per annum; this is recent geological past (Berger, 1988). ,2% of the global mean value in the contem- porary ocean.

8.05.13 IRON STIMULATION OF NUTRIENT UTILIZATION 8.05.14 LINKING IRON TO N2 FIXATION The enhancement of net export production Iron also appears to exert a strong constraint on (i.e., “true” new production) requires the addition N2 fixation in the modern ocean. Nitrogen-fixing of a limiting nutrient to the ocean, an increase in cyanobacteria, like most cyanobacteria, require the efficiency of utilization of preformed nutrients relatively high concentrations of iron (Berman- in the upper ocean, and/or a change in the Frank et al., 2001). The high-iron requirements elemental stoichiometry of primary producers come about because these organisms generally (Falkowski et al., 1998; Sarmiento and Bender, have high requirements for this element in their 1994). Indeed, an analysis of ice cores from photosynthetic apparatus (Fujita et al., 1990), as Antarctica, reconstruction of Aeolian iron deposi- well as for nitrogen fixation and electron carriers tions and concurrent atmospheric CO2 concen- that are critical for providing the reductants for 5 trations over the past 4.2 £ 10 yr (spanning four CO2 and N2 fixation in vivo. Increased Aeolian flux glacial–interglacial cycles) suggests that, when of iron to the oceans during glacial periods may iron fluxes were high, CO2 levels were low and have therefore not only stimulated the utilization vice versa (Martin, 1990; Petit et al., 1999). of nutrients in high-nutrient, low-chlorophyll Variations in iron fluxes were presumably a regions, but also stimulated nitrogen fixation by consequence of the areal extent of terrestrial cyanobacteria, and hence indirectly provided a deserts and wind vectors. It is hypothesized that significant source of new nitrogen. Both effects increased fluxes of iron to the high-nutrient, low would have led to increased photosynthetic carbon 208 Biogeochemistry of Primary Production in the Sea

fixation, and a net drawdown of atmospheric CO2 available in the oxidized ocean. For example, (Falkowski, 1997; Falkowski et al., 1998). a superoxide dismutase evolved in the green algae, and hence higher plants (and many non- photosynthetic eukaryotes) that utilized copper and zinc (Falkowski, 1997). Similar metal substitutions 8.05.15 OTHER TRACE-ELEMENT occurred in the photosynthetic apparatus and in CONTROLS ON NPP mitochondrial electron transport chains. Is iron the only limiting trace element in the The overall consequence of the co-evolution of sea? Prior to the evolution of oxygenic photosyn- oxygenic photosynthesis and the redox state of thesis, the oceans contained high concentrations the ocean is a relatively well-defined trace- (,1 mM) of dissolved iron in the form of Fe2þ element composition of the bulk phytoplankton. and manganese (.1 mM) in the form of Mn2þ, Analogous to Redfield’s relationship between the but essentially no copper or molybdenum; these macronutrients, trace-element analyses of phyto- and other elements would have been precipitated plankton reveals a relation for trace elements as sulfides. Thus, both iron and manganese normalized to cell phosphorus of (C125N16P1S1.3 were readily available to the early photoauto- K1.7Mg0.56Ca0.4)1000Sr4.4Fe7.5Mn3.8Zn0.80Cu0.38 trophs. The availability of these two elements Co0.19Cd0.21Mo0.03. The relative composition of permitted the evolution of the photosynthetic transition trace elements in phytoplankton is apparatus and the oxygen-evolving system that reflected in that of black shales (Figure 7). ultimately became the genetic template for all Thus, while there is a strong correlation bet- oxygenic photoautotrophs (Blankenship, 1992). ween N:P ratios in phytoplankton and seawater Indeed, the availability of these transition metals, (Anderson and Sarmiento, 1994; Lenton and which is largely determined by the oxidation state Watson, 2000), there is no parallel correlation of the environment, appears to account for their for trace elements (Whitfield, 2001; Morel and use in photosynthetic reactions. However, over Hudson, 1985)(Figure 7 inset). There are two geological time, oxygenic photoautotrophs them- underlying explanations for the lack of a strong selves altered the redox state of the ocean, and relationship. First, low abundance cations that hence the availability of the elements in the have a valance state of two or higher are often soluble phase (Anbar and Knoll, 2002). assimilated with particles, whether or not they are As photosynthetic oxygen evolution proceeded metabolically required. Hence, while zinc, 1 in the Proterozoic oceans, singlet oxygen ( O2), manganese, and cobalt are depleted in surface z2 peroxide (H2O2), superoxide anion radicals (O2 ), waters and are used in metabolic cycles, mercury, and hydroxide radicals (zOH) were all formed as lanthanum, and other rare-earth elements are also by-products (Kasting, 1990; Kasting et al., 1988). depleted and have no known biological function. These oxygen derivatives can oxidize proteins and The profiles of these elements are dictated, to first photosynthetic pigments as well as cause damage order, by particle fluxes and their ligands. Second, to reaction centers (Asada, 1994). A range of the absolute abundance of transition metals is molecules evolved to scavenge or quench the critically dependent on the solubility of the source potentially harmful oxygen by-products; these minerals, which is, in turn, regulated by redox z2 include superoxide dismutase (which converts O2 chemistry. Most key metabolic pathways evolved to O2 and H2O2), peroxidase (which reduces H2O2 prior to the oxidation of Earth’s atmosphere and to H2O by oxidizing an organic cosubstrate for the ocean, and hence many of the transition metals enzyme), and catalase (which converts 2H2O2 to selected for catalyzing biological redox reactions H2O and O2). The oldest superoxide dismutases reflect their relative abundance under anoxic or contained iron and/or manganese, while the suboxic conditions. For example, although the peroxidases and catalases contained iron (Asada contemporary ocean is oxidized, no known metal et al., 1980). These transition metals facilitate the can substitute for manganese in the water splitting electron transfer reactions that are at the core of complex in the photosynthetic apparatus. Simi- the respective enzyme activity, and their incor- larly, the photosynthetic machinery has maintain- poration into the proteins undoubtedly occurred ed a strict requirement for iron for over 2.8 Ga, because the metals were readily available and all nitrogenases require at least 16 iron (Williams and Frausto da Silva, 1996). As O2 atoms per enzyme complex. Some key biologi- production proceeded, the oxidation of Fe2þ, cal processes do not have the flexibility to Mn2þ, and S22 eventually led to the virtual substitute trace elements based simply on their depletion of these forms of the elements in the availability (Williams and Frausto da Silva, euphotic zone of the oceans. The depletion of 1996). Hence, while oxygenic photoautotrophs these elements had profound consequences on the indirectly determine the distribution of the major subsequent evolution of life. In the first instance, a trace elements in the ocean interior, the distri- number of enzymes were selected that incorpo- bution of these elements in the soluble phase does rated alternative transition metals that were not reflect with the composition of the organisms. References 209

105

4 10 Fe

103 Zn Mn

1 102 Cu Mo 101

Mo 2 Co 10 Zn Cu 10 103 Cd “Excess” elemental concentrations in black shales and sapropels (ppm) Cd Fe 4 Mn of elements (ppm) 10 Seawater concentrations Seawater 105 Co 1 1 10 100 1,000 1 10 100 1,000 Concentration of elements in phytoplankton (ppm)

Figure 7 The excess trace element composition in black shales compared with that of calculated marine phytoplankton, and the relationship between trace elements in seawater and phytoplankton (inset) (source Quigg et al., 2003).

On long timescales, seawater concentrations of radiation, a disequilibrium in geochemical pro- trace elements reflect the balance between their cesses, such that Earth maintains an oxidized sources and sinks. For trace elements, the sink atmosphere and ocean. This disequilibrium pre- term is tightly coupled with the extent of redox vents atmospheric oxygen from being depleted, conditions throughout the ocean, where periods of maintains a lowered atmospheric CO2 concen- extended reducing conditions result in greater tration, and simultaneously imprints on the ocean partitioning of organic carbon into the deep ocean interior and lithosphere elemental composition and sediments. This, in turn, leads to enhanced that reflect those of the bulk biological material sequestering of redox-sensitive trace elements into from which it is derived. sediments, thereby decreasing their seawater While primary producers in the ocean comprise concentrations. These sedimentary rocks have a only ,1% of Earth’s biomass, their metabolic rate high content of marine fractions (i.e., organic and biogeochemical impact rivals the much larger matter, apatite, biogenic silica and carbonates), terrestrial ecosystem. On geological timescales, and so are enriched by 1–2 orders of magnitude in these organisms are the little engines that are several trace elements: zinc, copper, nickel, moly- essential to maintaining life as we know it on bdenum, chromium, and vanadium (Piper, 1994). this planet. The positive correlation between trace-element ratios in phytoplankton and sediments is consist- ent with the notion that phytoplankton have imprinted their activities on the lithosphere. ACKNOWLEDGMENTS The author’s research is supported by grants from the US National Science Foundation, the National Aeronautical and Space Administration, 8.05.16 CONCLUDING REMARKS the US Department of Energy, and the US The evolution of primary producers in the Department of Defense. oceans profoundly changed the chemistry of the atmosphere, ocean, and lithosphere of Earth. The photosynthesis processes catalyzed by REFERENCES ensemble of these organisms not only influences the six major light elements, but directly and Abraham E. R., Law C. S., Boyd P. W., Lavender S. J., indirectly affect every major soluble redox- Maldonado M. T., and Bowie A. R. (2000) Importance of stirring in the development of an iron-fertilized phytoplank- sensitive trace element and transition metal on ton bloom. Nature 407(6805), 727–730. Earth’s surface. These processes continue to Aller R. C. (1998) Mobile deltaic and continental shelf muds as provide, primarily through the utilization of solar fluidized bed reactors. Mar. Chem. 61, 143–155. 210 Biogeochemistry of Primary Production in the Sea

Anbar A. D. and Knoll A. H. (2002) Proterozoic ocean Berman-Frank I., Cullen J. T., Shaked Y., Sherrell R. M., and chemistry and the evolution: a bioinorganic bridge? Science Falkowski P. G. (2001) Iron availability, cellular iron quotas, 297, 1137–1142. and nitrogen fixation in Trichodesmium. Limnol. Oceanogr. Anderson L. and Sarmiento J. (1994) Redfield ratios of 46, 1249–1260. remineralization determined by nutrient data analysis. Berner R. A. (1990) Atmospheric carbon dioxide levels over Global Biogeochem. Cycles 8, 65–80. phaneroic time. Science 249, 1382–1386. Antoine D., Andre J. M., and Morel A. (1996) Oceanic primary Berner R. A. and Maasch K. (1996) Chemical weathering production 2. Estimation at global-scale from satellite and controls on atmospheric O2 and CO2: fundamental (coastal zone color scanner) chlorophyll. Global Biogeo- principles were enunciated by J. J. Ebelmen in 1845. chem. Cycles 10, 57–69. Geochem. Geophys. Geosys. 60, 1633–1637. Asada K. (1994) Mechanisms for scavenging reactive mole- Berner R. A., Lasaga A., and Garrels R. (1983) The carbonate- cules generated in chloroplasts under light stress. In Photo- silicate geochemical cycle and its effect of atmospheric inhibition of Photosynthesis: from Molecular Mechanisms to carbon dioxide over the past 100 million years. Am. J. Sci. the Field (eds. N. Baker and J. Bowyer). Bios Scientific, 283, 641–683. Cambridge, pp. 129–142. Berthon J. F. and Morel A. (1992) Validation of a spectral Asada K., Kanematsu S., Okada S., and Hayakawa T. (1980) light-photosynthesis model and use of the model in Phytogenetic distribution of three types of superoxide conjunction with remotely sensed pigment observations. dismutases in organisms and cell organelles. In Chemical Limnol. Oceanogr. 37(4), 781–796. and Biochemical Aspects of Superoxide Dismutase (eds. Bienfang P. K. (1992) The role of coastal high latitude J. V. Bannister and H. A. O. Hill). Elsevier, Amsterdam, ecosystems in global export production. In Primary pp. 128–135. Productivity and Biogeochemical cycles in the Sea Azam F. (1998) Micriobial control of oceanic carbon flux: the (eds. P. G. Falkowski and A. Woodhead). Plenum, New plot thickens. Science 280, 694–696. York, pp. 285–297. Balch W. M., Holligan P. M., Ackleson S. G., and Voss K. J. Blackman F. F. (1905) Optima and limiting factors. Ann. Bot. (1991) Biological and optical properties of mesoscale 19, 281–298. coccolithophore blooms in the Gulf of Maine. Limnol. Blankenship R. E. (1992) Origin and early evolution of Oceanogr. 36(4), 629–643. photosynthesis. Photosyn. Res. 33, 91–111. Banse K. (1977) Determining the carbon-to-chlorophyll ratio Brasier M. D., Green O. R., Jephcoat A. P., Kleppe A. K., of natural phytoplankton. Mar. Biol. 41, 199–212. Van Kranendonk M. J., Lindsay J. F., Steele A., and Banse K. (1992) Grazing, temporal changes of phytoplankton Grassinau N. V. (2002) Questioning the evidence for Earth’s concentrations, and the microbial loop in the open sea. In oldest fossils. Nature 416, 76–81. Primary Productivity and Biogeochemical Cycles in the Bricaud A. and Morel A. (1987) Atmospheric corrections and interpretation of marine radiances in CZCS imagery: use of Sea (ed. P. G. Falkowski). Plenum, New York and London, a reflectance model. In Oceanography From Space: pp. 409–440. Proceedings Of The Atp Symposium On Remote Sensing, Banse K. (1994) Grazing and zooplankton production as key Brest, France vol. 7, pp. 33–50. controls of phytoplankton production in the open ocean. Broecker W. S., Peng T.-H., and Engh R. (1980) Modeling Oceanogr. 7, 13–20. the carbon system. Radiocarbon 22, 565–598. Barber R. T. (1992) Geological and climatic time scales of Brumsack H.-J. (1986) The inorganic geochemistry of Cretac- nutrient availability. In Primary Productivity and Biogeo- eous black shales (DSDP Leg 41) in comparison to modern chemical Cycles in the Sea (eds. P. G. Falkowski and upwelling sediments from the Gulf of California. In A. Woodhead). Plenum, New York, pp. 89–106. North Atlantic Palaeoceanography, Geological Society Barron E. J., Fawcett P. J., Peterson W. H., Pollard D., and Special Publication. No. 21, (eds. C. P. Summerhayes and Thompson S. L. (1995) A simulation of midcretaceous N. J. Shackleton), pp. 447–462. climate. Paleoceanography 10(5), 953–962. Campbell J. W. and Aarup T. (1992) New production in the Beaumont V. I., Jahnke L. L., and Des Marais D. J. (2000) North Atlantic derived from the seasonal patterns of surface Nitrogen isotopic fractionation in the synthesis of photo- chlorophyll. Deep-Sea Res. 39, 1669–1694. synthetic pigments in Rhodobacter capsulatus and Anabaena Capone D. G. and Carpenter E. J. (1982) Nitrogen fix- cylindrica. Org. Geochem. 31(11), 1075–1085. ation in the marine environment. Science 217(4565), Behrenfeld M. J. and Falkowski P. G. (1997a) Photosynthetic 1140–1142. rates derived from satellite-based chlorophyll concentration. Chandler M. A., Rind D., and Ruedy R. (1992) Pangean Limnol. Oceanogr. 42, 1–20. climate during the Early Jurassic: GMC simulations and the Behrenfeld M. and Falkowski P. (1997b) A consumer’s guide sedimentary record of paleoclimate. Geol. Soc. Am. Bul. 104, to phytoplankton productivity models. Limnol. Oceanogr. 543–559. 42, 1479–1491. Christensen J. P., Murray J. W., Devol A. H., and Codispoti Behrenfeld M., Bale A., Kolber Z., Aiken J., and Falkowski P. L. A. (1987) Denitrification in continental shelf sediments (1996) Confirmation of iron limitation of phytoplankton has major impact on the oceanic nitrogen budget. Global photosynthesis in the equatorial Pacific. Nature 383, Biogeochem. Cycles 1, 97–116. 508–511. Codispoti L. (1995) Is the ocean losing nitrate? Nature 376, Beja O., Spudich E. N., Spudich J. L., Leclerc M., and DeLong 724. E. F. (2001) Proteorhodopsin phototrophy in the ocean. Codispoti L. A. and Christensen J. P. (1985) Nitrification, Nature 411(6839), 786–789. denitrification and nitrous oxide cycling in the eastern Berger A. (1988) Milankovitch theory and climate. Rev. tropical south Pacific Ocean. Mar. Chem. 16, 277–300. Geophys. 26, 624–657. Conley D. J. (2002) Terrestrial ecosystems and the global Berger W. H., Smetacek V. S., et al. (eds.) (1989) Productivity biogeochemical cycle. Glob. Biogeochem. Cycles 16, doi: of the Ocean: Present and Past. Wiley, New York, 471pp. 10.129/2002GB001894. Berggren W. A., Kent D. V., Swisher C. C., and Aubry M.-P. Copin-Montegut C. and Copin-Montegut G. (1983) Stoi- (1995) A revised Cenozoic geochronology and chronostrati- chiometry of carbon, nitrogen, and phosporus in marine graphy. In Geochronology, Time Scales, and Global particulate matter. Deep-Sea Res. 30, 31–46. Stratigraphic Correlations: A Unified Temporal Framework Delwiche C. (2000) Tracing the thread of plastid diversity for an Historical Geology, Spec. Vol. Soc. Econ. Paleontol. through the tapestry of life. Am. Nat. 154, S164–S177. Mineral. 54 (eds. W. A. Berggren, D. V. Kent, and Des Marais D. J. (2000) When did photosynthesis emerge on J. Hardenbol). pp. 129–212. Earth? Science 289, 1703–1705. References 211

Devol A. H. (1991) Direct measurement of nitrogen Habicht K. S., Gade M., Thamdrup B., Berg P., and Canfield gas fluxes from continental shelf sediments. Nature 349, D. E. (2002) Calibration of sulfate levels in the Archean 319–321. Ocean. Science 298, 2372–2374. Duce R. A. and Tindale N. W. (1991) Atmospheric transport of Harrison W. G. (1980) Nutrient regeneration and primary iron and its deposition in the ocean. Limnol. Oceanogr. 36, production in the sea. In Primary Production in the Sea (ed. 1715–1726. P. G. Falkowski). Plenum, New York, pp. 433–460. Dugdale R. C. (1967) Nutrient limitation in the sea: dynamics, Harwood D. M. and Nikolaev V. A. (1995) Cretaceous diatoms: identification and significance. Limnol. Oceanogr. 12, morphology, taxonomy, biostratigraphy. In Siliceous Micro- 685–695. fossils (eds. C. D. Blome, P. M. Whalen, and R. Katherine). Dugdale R. C. and Goering J. J. (1967) Uptake of new and Paleontological Society, Number 8, pp. 81–106. regenerated forms of nitrogen in primary productivity. Herbert D., Elsworth R., and Telling R. C. (1956) The Limnol. Oceanogr. 12, 196–206. continuous culture of bacterial a theoretical and experimen- Dugdale R. and Wilkerson F. (1992) Nutrient limitation of new tal study. J. Gen. Microbiol. 14, 601–622. production in the sea. In Primary Productivity and Holland H. and Rye R. (1998) Am. J. Sci. 298, 621–672. Biogeochemical Cycles in the Sea. (eds. P. G. Falkowski). Holligan P. M. and Balch W. M. (1991) From the ocean to Plenum, New York and London, pp. 107–122. cells: coccolithophore optics and biogeochemistry. In Einstein A. (1910) Therioe der Opaleszenz von homogenen Particle Analysis in Oceanography 27 (eds. S. Demers). Flussigkeiten un Flussigkeitsgemischen in der Nahe des Springer, Berlin, pp. 301–324. kritischen Zustandes. Ann. Physik. 33, 1275. Huber B. T., Hodell D., and Hamilton C. (1995) Middle–Late Eppley R. W. (1992) Towards understanding the roles of Cretaceous climate of the southern high Latitudes—stable phytoplankton in biogeochemical cycles: personal notes. In isotopic evidence for minimal equator-to-pole thermal Primary Productivity and Biogeochemical Cycles in the gradients. Geol. Soc. Am. Bull. 107, 1164–1191. Sea (eds. P. G. Falkowski and A. D. Woodhead). Plenum, Iglesias-Rodriguez D. M., Brown C. W., Doney S. C., Kleypas New York and London, pp. 1–7. J. A., Kolber D., Kolber Z., Hayes P. K., and Falkowski G. P. Eppley R. W. and Peterson B. J. (1979) Particulate organic (2002) Representing key phytoplankton functional groups in matter flux and planktonic new production in the deep ocean. ocean carbon cycle models: coccolithophorids. Global Nature 282, 677–680. Biogeochem. Cycles 16 (in press). Eppley R. W., Rogers J. N., and McCarthy J. J. (1969) Janis N. and Damuth J. (1990) Mammals. In Evolutionary Half-Saturation constant for uptake of nitrate and ammo- Trends (ed. K. McNammara). Belknap, London, pp. nium by marine phytoplankton. Limnol. Oceanogr. 14, 301–345. 912–920. Johnston A. M. and Raven J. A. (1987) The C{-4}- Falkowski P. (1997) Evolution of the nitrogen cycle and its like characteristics of intertidal macroalga Ascophyllum influence on the biological sequestration of CO in the 2 nedosum (Fucales, Phaeophyta). Phycologia 26(2), ocean. Nature 387, 272–275. 159–166. Falkowski P. G. (2002) On the evolution of the carbon cycle. Kasting J. F. (1990) Bolide impacts and the oxidation state of In Phytoplankton Productivity: Carbon Assimilation in carbon in the Earth’s early atmosphere. Origins Life Evol. Marine and Freshwater Ecosystems (eds. P. I. B. Williams, Biosphere 20, 199–231. D. Thomas, and C. Renyolds). Blackwell, Oxford, Kasting J. F. (1993) Earth’s early atmosphere. Science 259, pp. 318–349. 920–926. Falkowski P. G. and Raven J. A. (1997) Aquatic Photosyn- Kasting J. F., Toon O. B., and Pollack J. B. (1988) How climate thesis. Blackwell, Oxford, 375pp. evolved on the terrestrial planets. Sci. Am. 258, 90–97. Falkowski P., Barber R., and Smetacek V. (1998) Biogeochemi- cal controls and feedbacks on ocean primary production. Katz M. E., Wright J. D., Miller K. G., Cramer B. S., Fennel K., Science 281, 200–206. and Falkowski P. G. Biological overprint of the geological Fanning K. (1992) Nutrient provinces in the sea: concentration carbon cycle. Nature (in review). ratios, reaction rate ratios, and ideal covariation. J. Geophys. Kolber Z. S., Barber R. T., Coale K. H., Fitzwater S. E., Res. 97C, 5693–5712. Greene R. M., Johnson K. S., Lindley S., and Falkowski Farquhar J., Wing B. A., McKeegan K. D., Harris J. W., P. G. (1994) Iron limitation of phytoplankton photosynthesis Cartigny P., and Thiemens M. H. (2002) Mass-independent in the Equatorial Pacific Ocean. Nature 371, 145–149. sulfur of inclusions in diamond and sulfur recycling on early Kromer S. (1995) Respiration during photosynthesis. Ann. Rev. Earth. Science 298(5602), 2369–2372. Plant Physiol. Plant Mol. Biol. 0046, 00045–00070. Field C., Behrenfeld M., Randerson J., and Falkowski P. (1998) Kump L. and Arthur M. (1999) Interpreting carbon-isotope Primary production of the biosphere: integrating terrestrial excursions: carbonates and organic matter. Chem. Geol. 161, and oceanic components. Science 281, 237–240. 181–198. Fujita Y., Murakami A., and Ohki K. (1990) Regulation of the Laws E. A., Popp B. N., and Bidigare R. (1997) Effect of growth stoichiometry of thylakoid components in the photosynthetic rate and CO2 concentration on carbon siotoic fractionation system of cyanophytes: model experiments showing that by the marine diatom Phaeodactylum tricornutum. Limnol. control of the synthesis or supply of Ch A can change the Oceanogr. (in press). stoichiometric relationship between the two photosystems. Lenton T. and Watson A. (2000) Redfield revisited: 2. What Plant. Cell. Physiol. 31, 145–153. regulates the oxygen content of the atmosphere. Global Gordon H. R. and Morel A. (1983) Remote Sensing of Ocean Biogeochem. Cycles 14, 249–268. Color for Interpretation of Satellie Visible Imagery: A Letelier R. M., Dore J. E., Winn C. D., and Karl D. M. (1996) Review. Springer, New York. Seasonal and interannual variations in photosynthetic carbon Gradstein F. M., Agterberg F. P., Ogg J. G., Hardenbol H., van assimilation at Station ALOHA. Deep-Sea Res. 43, Veen P., Thierry J., and Huang Z. A. (1995) A Triassic, 467–490. Jurassic, and Cretaceous time scale. In Geochronology, Liebig J. (1840) Chemistry and its Application to Agriculture Time Scales, and Global Stratigraphic Correlations: A and Physiology. Taylor and Walton, London. Unified Temporal Framework for an Historical Geology Lindeman R. (1942) The trophic-dynamic aspect of ecology. (eds. W. A. Berggren, D. V. Kent, and J. Hardenbol). Spec. Ecology 23, 399–418. Vol.- Soc. Econ. Paleontol. Mineral. 54, 95–126. Lipps J. H. (ed.) (1993) Fossil Prokaryotes and Protists. Gruber N. and Sarmiento J. (1997) Global patterns of marine Blackwell, Oxford. nitrogen fixation and denitrification. Global Biogeochem. Martin J. H. (1990) Glacial-interglacial CO2 change: the iron Cycles 11, 235–266. hypothesis. Paleoceangraphy 5, 1–13. 212 Biogeochemistry of Primary Production in the Sea

Martin J. H. (1991) Iron, Liebig’s Law and the greenhouse. (ed. P. G. Falkowski). Plenum, New York and London, Oceanography 4, 52–55. pp. 239–255. McElroy M. (1983) Marine biological controls on atmospheric Raven J. A. (1987) The role of vacuoles. New Phytol. 106, CO2 and climate. Nature 302, 328–329. 357–422. Miller C. B., Frost B. W., Wheeler P. A., Landry M. R., Raymo M. and Ruddiman W. (1992) Tectonic forcing of the Welschmeyer N., and Powell T. M. (1991) Ecological late Cenozoic climate. Nature 359, 117–122. dynamics in the subarctic Pacific, a possibly iron-limited Redfield A. C. (1934) On the Proportions of Organic Deri- ecosystem. Limnol. Oceanogr. 36, 1600–1615. vatives in Sea Water and Their Relation to the Com- Milliman J. (1993) Production and accumulation of calcium position of Plankton. Liverpool, James Johnstone Memorial carbonate in the ocean: budget of a nonsteady state. Global Volume, Liverpool Univ. Press, pp. 176–192. Biogeochem. Cycles 7, 927–957. Redfield A. C. (1958) The biological control of chemical Monod J. (1942) Recheres sur la croissance des cultures factors in the environment. Am. Sci. 46, 205–221. bacteriennes. Hermann & Cie, Paris. Redfield A., Ketchum B., and Richards F. (1963) The influence Morel A. (1974) Optical properties of pure water and of organisms on the composition of sea-water. In The Sea pure seawater. In Optical Aspects of Oceanography (eds. (ed. M. Hill). Interscience, New York, 2, 26–77. N. G. Jerlov and E. S. Nielsen). Academic Press, London, pp. Retallack G. (2001) Cenozoic expansion of grasslands and 1–24. climatic cooling. J. Geol. 109, 407–426. Morel A. (1988) Optical modeling of the upper ocean in Sancetta C., Villareal T., and Falkowski P. G. (1991) Massive relation to it’s biogenous matter content (case one waters). fluxes of rhizosolenoid diatoms: a common occurrence? J. Geophys. Res. 93, 10749–10768. Limnol. Oceanogr. 36, 1452–1457. Morel F. M. M. and Hudson R. J. (1985) The geobiological Sarmiento J. L. and Bender M. (1994) Carbon biogeochemistry cycle of trace elements in aquatic systems: Redfield and climate change. Photosyn. Res. 39, 209–234. revisited. In Chemical Processes in Lakes (ed. W. Strumm). Sarmiento J. L., Orr J. C., and Siegenthaler U. (1992) Wiley-Interscience, pp. 251–281. A perturbation simulation of CO2 uptake in an ocean general Morel A. and Andre J. M. (1991) Pigment distribution and circulation model. J. Geophys. Res. 94, 3621–3645. primary production in the western Mediterranean as derived Schlesinger W. H. (1997) Biogeochemistry: An Analysis of and modeled from coastal zone color scanner observations. Global Change. Academic Press, New York. J. Geophys. Res. C. Oceans 96(C7), 12685–12698. Schopf J. (1993) Microfossils of the early Archean Apex Mulkidjanian A. and Junge W. (1997) On the origin of Chert: new evidence of the antiquity of life. Science 260, photosynthesis as inferred from sequence analysis—a 640–646. primordial UV-protector as common ancestor of reaction Siegenthaler U. and Sarmiento J. L. (1993) Atmospheric carbon Nature 365 centers and antenna proteins. Photosyn. Res. 51, 27–42. dioxide and the ocean. , 119–125. Sigman D. and Boyle E. (2000) Glacial/interglacial variations Nixon S. W., Ammerman J. W., Atkinson L. P., Berounsky in atmospheric carbon dioxide. Nature 407, 859–869. V. M., Billen G., Boicourt W. C., Boynton W. R., Smetacek V. (1999) Diatoms and the ocean carbon cycle. Church T. M., Ditoro D. M., Elmgren R., Garber J. H., Protist 150, 25–32. Giblin A. E., Jahnke R. A., Owens N. J. P., Pilson M. E. Q., Smith D. F. (1981) Tracer kinetic analysis applied to problems and Seitzinger S. P. (1996) The fate of nitrogen and in marine biology. In Physiological Bases of Phytoplankton phosphorus at the land-sea margin of the North Atlantic Ecology (ed. T. Platt) Can. Bull. Fish Aquat. Sci. 170, Ocean. Biogeochemistry 35, 141–180. Ottawa, pp. 113–129. Odum E. P. (1971) Fundamentals of Ecology. Philadelphia. Smith W. O., Jr. and Nelson D. M. (1990) The importance of Pace N. R. (1997) A molecular view of microbial diversity and ice-edge phytoplankton production in the Southern Ocean. the biosphere. Science 276(5313), 734–740. BioScience 36, 151–157. Pavlov A. and Kasting J. (2002) Mass-independent fraction- Stumm W. and Morgan J. J. (1981) Aquatic Chemistry. Wiley, ation of sulfur isotopes in Archean sediments: strong New York. evidence for an anoxic Archean atmosphere. Astrobiology Summons R., Jahnke L., Hope J., and Logan G. (1999) 2, 27–41. 2-Methylhopanoids as biomarkers for cyanobacterial Petit J. R., Jouzel J., Raynaud D., Barkov N. I., Barnola J.-M., oxygenic photosynthesis. Nature 400, 55–557. and Basile I. (1999) Climate and atmospheric history of the Tilman D. (1982) Resource Competition and Community past 420,000 years from the Vostok ice core, Antarctica. Structure. Princeton University Press, Princeton. Nature 399, 429–436. Tozzi S. (2001) Competition and succession of key marine Pickard G. L. and Emery W. J. (1990) Descriptive Physical phytoplankton functional groups in a variable environment. Oceanography. Pergamon, Oxford. IMCS. New Brunswick, Rutgers University, 79. Piper D. Z. (1994) Seawater as the source of minor elements in Tozzi S., Schofield O., and Falkowski P. (2003) Turbulence as black shales, phosphorites and other sedimentary rocks. a selective agent of two phytoplankton functional groups. Chem. Geol. 114, 95–114. Global Change Biology (in press). Platt T. (1986) Primary production of the ocean water column Tsuda A. and Kawaguchi S., et al. (2003) Microzooplankton as a function of surface light intensity: algorithms for remote grazing in the surface-water of the Southern Ocean during an sensing. Deep-Sea Res. 33, 149–163. Austral Summer. Polar Biology. Platt T. and Irwin B. (1973) Caloric content of phytoplankton. Tyrell T. (1999) The relative influences of nitrogen and Limnol. Oceanogr. 18, 306–310. phosphorus on oceanic primary production. Nature 400, Platt T. and Sathyendranath S. (1988) Oceanic primary 525–531. production: estimation by remote sensing at local and Veizer J., Ala D., Azmy K., Bruckschen P., Buhl D., Bruhn F., regional scales. Science 241, 1613–1620. Carden G. A. F., Diener A., Ebneth S., Godderis Y., Jasper Postgate J. R. (ed.) (1971) The Chemistry and Biochemistry of T., Korte C., Pawellek F., Podlaha O. G., and Strauss H. Nitrogen Fixation. Plenum, New York. (1999) 87Sr/86Sr, d13C and d18O evolution of Phanerozoic Quigg A., Finkel Z. V., Irwin A. J., Rosenthal Y., Ho T.-Y., seawater. Chem. Geol. 161, 59–88. Reinfelder J. R., Schofield O., More F., and Falkowski P. Volk T. and Hoffert M. I. (1985) Ocean carbon pumps: analysis (2003) Plastid inheritance of elemental stoichiometry in of relative strengths and efficiencies in ocean-driven phytoplankton and its imprint on the geological record. atmospheric CO2 exchanges. In The Carbon Cycle and Nature 425, 291–294. Atmospheric CO2: Natural Variations Archean to Present Ramus J. (1992) Productivity of seaweeds. In Primary (eds.E.T.SunquistandW.S.Broeker).American Productivity and Biogeochemical Cycles in the Sea Geophysical Union. 32, Washington, DC, pp. 99–110. References 213

Whitfield M. (2001) Interactions between phytoplankton and Wu J., Sunda W., Boyle E. A., and Karl D. M. (2000) trace metals in the ocean. Adv. Mar. Biol. 41, 3–128. Phosphate depletion in the western North Atlantic ocean. Walsh J. J. (1983) Death in the sea: enigmatic phytoplankton Science 289(5480), 759–762. losses. Prog. Oceanogr. 12, 1–86. Xiong J., Fischer W. M., Inoue K., Nakahara M., and Warneck P. (1988) Chemistry of the Natural Atmosphere. Bauer C. E. (2000). Molecular evidence for the early Academic Press, New York. evolution of photosynthesis. Science 289(5485), Williams P. J. L. (1981) Incorporation of microheterotrophic 1724–1730. processes into the classical paradigm of the planktonic Yoder J. A., McClain C. R., Feldman G. C., Esaias W. E., food web. Kieler Meeresforsch. Sonderh. 5, 1–27. (1993) Annual cycles of phytoplankton chlorophyll concen- Williams P. J. L. (1993) On the definition of plankton trations in the global ocean: a satellite view. Global production terms. ICES Mar. Sci. Symp. 197, 9–19. Biogeochem. Cycles 7, 181–193. Williams R. and Frausto da Silva J. (1996) The Natural Selection Zehr J. P., Mellon M., Braun S., Litaker W., Steppe T., of the Chemical Elements. Clarendon Press, Oxford. and Paerl H. W. (1995) Diversity of hetero- Woodwell G. M., Whittaker R. H., Reiners W. A., Likens G. E., trophic nitrogen-fixation genes in a marine cyano- Delwiche C. C., and Botkin D. B. (1978) The biota and the bacterial mat. Appl. Environ. Microbiol. 0061, world carbon budget. Science 199, 141–146. 02527–02532.

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 8; (ISBN: 0-08-044343-5); pp. 185–213