IODP

doi:10.2204/iodp.sp.329.2010

Scientific objectives

Our study has the following fundamental objectives:

  1. To document the habitats, metabolic activities, genetic composition, and biomass of microbial communities in subseafloor sediment with very low total activity.
  2. To test how oceanographic factors (such as surface ocean productivity) control variation in sedimentary habitats, activities, and communities from gyre center to gyre margin.
  3. To quantify the extent to which these sedimentary communities may be supplied with electron donors (food) by water radiolysis, a process independent of the surface photosynthetic world.
  4. To determine how the basement habitats, potential activities, and, if measurable, communities vary with crust age and hydrologic regime (from ridge crest to abyssal plain).

We propose to meet these objectives by (1) coring the entire sediment column at several sites along two transects in the region of the South Pacific Gyre (Figs. F1, F2); (2) coring and logging the upper 100 m of the basaltic basement at three key sites; and (3) undertaking extensive microbiological, biogeochemical, geological, and geophysical analysis of the cores and drill holes.

The project results will address several significant questions. Are the communities in mid-gyre subseafloor sediment uniquely structured? Do they contain previously unknown organisms? What are their principal sources of metabolic energy? Do their principal activities and composition vary with properties of the surface world, such as sea-surface chlorophyll concentrations or organic flux to the seafloor? Is microbial activity is sustainable in subseafloor basalt by mineral oxidation (e.g., oxidation of iron in the basaltic minerals) for tens of millions of years after basalt formation? Are microbial communities recognizably present in subseafloor basalt older than 13 Ma?

These questions can be framed as hypotheses to be tested. For example, we hypothesize that

  1. A living community persists in the most organic poor sediment of the world ocean.
  2. Organic-fueled metabolic activity is extremely low and oxygen is the principal net terminal electron acceptor in this sediment. Consequently, the diversity of anaerobic activities is far lower here than in previously examined subseafloor sediment.
  3. The biomass and composition of this subseafloor sedimentary community is distinctly different from the communities observed in the higher activity anaerobic subseafloor ecosystems that have been examined to date.
  4. H2 from water radiolysis is a significant electron donor for microbes in the most organic-poor subseafloor sediment.
  5. Open flow continues in the very old basalt of the western gyre.
  6. Basalt oxidation may support microbial activity for 100 m.y. here.
  7. Biomass and activity decrease with basement age as electron donor accessibility decreases.

Even if none of the above hypotheses are found to be valid, the results of the proposed expedition will significantly advance understanding of the subsurface world. Scientists will test the extent to which distinct oceanographic and geologic provinces contain distinct subseafloor habitats and distinct subseafloor communities. They will document the extent to which life in the low-activity gyre sediment depends on the surface photosynthetic world—and the extent, if any, to which it is metabolically independent of the surface world. They will place firm constraints on the potential for microbial redox activity in ancient subseafloor basalt and how that potential varies with crust age over 100 m.y. or more. They will place firm constraints on estimates of total subseafloor biomass.

The results of the proposed expedition will also test the factors that control evolution of geothermal circulation and chemical alteration in ocean crust, models of regional tectonic history, geodynamo models, and models of glacial–interglacial ocean-climate change.

Explanation of primary objectives

1. To document the habitats, metabolic activities, genetic composition, and biomass of microbial communities in subseafloor sediment with very low total activity.

What are the principal microbial activities in mid-gyre subseafloor sediment? What are the rates of those activities? How dense are the populations in this sediment? What are the communities active here? How do these communities and activities compare to those in subseafloor sediment with much higher levels of activity? How unique are their organisms?

IODP drilling in the South Pacific Gyre will provide a crucial opportunity to document microbial habitats, activities, and community composition in a subseafloor sedimentary ecosystem that is unlike any that IODP or ODP has ever explored.

Metabolic activities

The penetration of high concentrations of O2 and NO3 meters into the sediment at proposed Sites SPG-1A–SPG-10A (Fig. F4) suggests that the net rate of electron-accepting activity is even lower in the South Pacific Gyre than at Site 1231 (Fig. F2). Along with chemical data from Deep Sea Drilling Project (DSDP) Leg 92, these O2 and NO3 profiles also suggest that the principal electron-accepting activity in the deep subseafloor sediment of this region may be different than at any of the sites where subseafloor sedimentary communities have been previously explored. A sequence of sites was drilled during Leg 92 at ~20°S in the northern portion of the gyre (Leinen, Rea, et al., 1986). At all Leg 92 sites where pore water chemistry was analyzed, dissolved NO3 is present throughout the entire sediment column, which is as thick as 50 m (Gieskes and Boulègue, 1986). Because NO3 concentrations do not significantly change with depth in the Cruise KNOX-02RR survey cores or the Leg 92 cores, O2 may be the principal net electron acceptor in subseafloor sediment throughout the South Pacific Gyre. For thermodynamic reasons, iron reduction and SO42– reduction cannot yield more energy than NO3 reduction at the NO3 concentrations of our surveyed sediment and the Leg 92 sites. Consequently, they are very unlikely to be significant electron acceptors in the sediment of this region.

Biomass

Quantification of subseafloor biomass in the South Pacific Gyre will place strong constraints on the size of Earth's subseafloor biomass and total biomass.

On the basis of acridine orange direct counts (AODC) in subseafloor sediment of relatively high activity sites, subseafloor sedimentary biomass has been estimated to constitute 1/10 to 1/3 of Earth's total living biomass (Whitman et al., 1998; Parkes et al., 2000). These estimates make two key assumptions that need to be tested in a range of different geologic settings. First, they assume that the sites where cells were counted are representative of biomass content in the subseafloor sediment of all regions. Second, they assume that all of the cells observed in AODC are alive. Because acridine orange and other deoxyribonucleic acid (DNA) stains (e.g, SybrGreen) bind to DNA and fluoresce whether cells are alive or dead, current estimates of subseafloor biomass place an upper bound on estimates of living biomass.

The results of Cruise KNOX-02RR indicate that the first assumption is wrong because previous sites have not sampled sediment in the major ocean gyres. As described earlier, cell concentrations in the Cruise KNOX-02RR cores are orders of magnitude lower than concentrations at the same depths in all previous IODP/ODP sediment (D'Hondt et al., 2009) (Fig. F3).

The fraction of cells alive and active in South Pacific Gyre sediment is not yet known. Catalyzed reporter deposition fluorescence in situ hybridization (CARD-FISH) studies demonstrate that 10% (or more) of the cells in Leg 201 subseafloor sediment are alive and active (Mauclaire et al., 2005). CARD-FISH differentiates active, ribosomal ribonucleic acid (rRNA)-rich cells from rRNA-depleted, inactive, or dead cells by targeting rRNA molecules that occur in high copy number in actively metabolizing cells. If only 10% of subseafloor sedimentary cells are active throughout the world and deep subseafloor gyre sediment contains as few cells per cubic centimeter as in the Cruise KNOX-02RR cores, living subseafloor biomass is much more than a factor of 10 lower than 10%–30% of Earth's total biomass (Kallmeyer et al., 2009).

Community composition

Genetic analyses of bulk sediment samples and cultured bacterial isolates from deep beneath the seafloor demonstrate that similar subseafloor sedimentary environments separated by thousands of kilometers contain similar genetic communities. For example, hydrate-bearing sediment of the Peru margin and the northwestern U.S. margin (Hydrate Ridge) contains similar 16S rRNA gene populations (Inagaki et al., 2006). In contrast, nearby sites with different sedimentary environments contain very different populations (Inagaki et al., 2006).

Genetic analyses of bulk sediment samples and cultured bacterial isolates also demonstrate that previously undiscovered lineages exist in subseafloor sediment, even within a few meters of the seafloor (Sørensen et al., 2004; D'Hondt et al., 2004). For example, 16S rRNA gene sequencing surveys of bulk sediment from a few meters beneath the seafloor at Site 1231 discovered previously unknown archaea with 16S sequences that are deeply rooted in the tree of life (Sørensen et al., 2004).

In combination with the metabolic points described above, these discoveries underscore three important points about potential community composition in subseafloor sediment of the South Pacific Gyre. First, the gyre's subseafloor microbial communities differ from those of any subseafloor sediment explored to date. Second, the very low activity and potentially aerobic communities of the deepest sediment in the South Pacific Gyre are likely to contain unique and previously undiscovered microorganisms. Third, exploration of the gyre sediment may provide deep insight into community composition and structure throughout most of the open ocean, because the genetic communities of subseafloor sediment in the other major ocean gyres may resemble those of the South Pacific Gyre.

2. To test how oceanographic factors (such as surface ocean productivity) control variation in sedimentary habitats, activities, and communities from gyre center to gyre margin.

How are subseafloor sedimentary activities and communities affected by oceanographic properties that vary predictably from gyre center to gyre margin?

The proposed expedition will provide a unique opportunity to document how the nature of subseafloor sedimentary life varies with oceanographic properties in the least biologically active region of the world ocean.

Pore water surveys of regions with higher organic fluxes to the seafloor, such as the equatorial Pacific and continental margins, indicate that subseafloor cell concentrations and the principal electron-accepting activity and the net rates of electron-accepting activities vary predictably with sea-surface chlorophyll content and organic flux to the seafloor (D'Hondt et al., 2002, 2004). These relationships appear principally due to reliance of subseafloor sedimentary communities on burial of photosynthesized organic matter for electron donors. Other sedimentary properties that vary with measures of oceanic productivity include sediment composition, which depends on the rate of microfossil production in the overlying water column as well as on the position of the carbonate compensation depth (CCD), and sediment thickness, which largely depends on the rate of calcium carbonate and biosilica production in the overlying water column.

The Cruise KNOX-02RR results (D'Hondt et al., 2009) suggest that the principal electron-accepting activity, net rates of activities, and cell concentrations in the shallow subseafloor sediment of the South Pacific Gyre vary with sea-surface chlorophyll content from within the gyre to outside the gyre and perhaps from site to site within it. If these relationships are shown by IODP drilling to also apply to the deeper subseafloor sediment of this region, then biomass, principal redox activities, and net rates of activity in subseafloor sediment may be largely predicted by surface ocean properties.

3. To quantify the extent to which these sedimentary communities may be supplied with electron donors (food) by water radiolysis.

To what extent does the ecosystem of organic-poor sediment depend on in situ radiolysis of pore water?

Expedition 329 will provide an unprecedented opportunity to determine if subseafloor life in very low activity sediment is nourished to a significant extent by H2 from in situ radiolysis of water.

Buried organic matter appears to be the principal source of electron donors in subseafloor sediment of ocean margins and the equatorial Pacific (D'Hondt et al., 2004). However, the burial flux of organic carbon is so low in the South Pacific Gyre that in situ radiolysis of water may be the principal source of electron donors there (D'Hondt et al., 2009) (Fig. F5).

Water radiolysis has been described as a possible source of energy for ecosystems in hard rock far beneath continental surfaces (Pedersen, 1996; Lin et al., 2005a). The longest-lived products of water radiolysis are the electron donor H2 and the electron acceptor O2 (Debierne, 1909). H2 can be supplied by in situ radiolysis or, in theory, by transport of radiolytic H2 from a much deeper biologically dead environment (e.g., the mantle and oceanic basement deep beneath sediment).

The potential importance of water radiolysis to subseafloor sedimentary communities can be assessed by comparing radiolytic H2 production to organic-fueled respiration. For this comparison, in situ H2 production by water radiolysis must be quantified from (1) logging estimates of uranium, thorium, and potassium concentrations (D'Hondt, Jørgensen, Miller, et al., 2003) or inductively coupled plasma–mass spectrometry (ICP-MS) data (Blair et al., 2007), (2) shipboard physical property measurements (porosity and density) (D'Hondt, Jørgensen, Miller, et al., 2003), and (3) a numerical model of water radiolysis (Blair et al., 2007).

Water radiolysis may provide a higher flux of electron donors than buried organic matter in gyre sediment. Organic-fueled respiration rates are much lower in this sediment than in any previously drilled regions of the world ocean. Clay-rich sediment contains much higher concentrations of radioactive elements than other deep-sea sediment. Furthermore, in fine-grained porous sediment, most alpha and beta production occurs within striking range of pore water. Consequently, clay-rich sediment will yield much higher rates of water radiolysis than other geological environments.

4. To determine how basement habitats, potential activities, and, if measurable, communities vary with crust age and hydrologic regime (from ridge crest to abyssal plain).

How does the habitability of subseafloor basalt change with crust age? How does this change depend on basement hydrologic evolution and mineral alteration? Do fractures remain open for flow-in basalt as old as 100 Ma? What is the role of sediment cover in controlling hydrologic flow, alteration, and habitability in subseafloor basalt? Are dissolved oxidants (oxygen and nitrate) available in thinly sedimented subseafloor basalt as old as 100 Ma? Are reduced elements (e.g., iron and sulfur) available for oxidation in basalt as old as 100 Ma? Does mineral oxidation support measurable microbial activity in this basalt?

The proposed expedition will provide a unique opportunity to determine how basement habitability and communities vary with crust age, sediment cover, and hydrologic conditions over 100 m.y. or more of basement history in the most thinly sedimented region of the world ocean.

Subseafloor basalt has been described as the largest potential microbial habitat on Earth (Fisk et al., 1998). Glass alteration textures have been interpreted as evidence of microbial colonization in subseafloor basalt as old as 145 Ma (Fisk et al., 1998; Furnes and Staudigel, 1999). Laboratory experiments indicate that microbial communities may play an important role in basalt weathering (Staudigel et al., 1998). Indeed, that microbes can derive their energy from rock weathering (e.g., oxidation of iron and sulfur in minerals) has been demonstrated with seafloor cultured isolates (Edwards et al., 2003). Fluid samples from a CORKed borehole show that microbes exist in the formation fluid of 3.5 Ma subseafloor basalt (Cowen et al., 2003).

Despite this evidence, the nature and extent of subseafloor basaltic communities remain largely unknown and the factors that control the "redox habitability" of the basalt (the ability of the basalt to fuel microbial reactions) are relatively unexplored. The biological significance of the textural features and the factors that control alteration and its timing in crust are not well constrained.

Metabolic habitability of subseafloor basalt ultimately depends on the supply of electron donors (principally, reduced Fe and S) from minerals in the basalt and the supply of electron acceptors (O2 and NO3) from seawater flowing into and through the basalt. Processes that change these supplies over the basalt's lifetime include mineral alteration, which changes the supply of electron donors, and the evolution of hydrologic flow through the basalt, which changes the supply of electron acceptors.

Evolution of crust alteration and metabolic habitability

Compilation of DSDP/ODP geochemical data, mostly from the North Atlantic, suggests that oxidation of Fe and S in the upper few hundred meters of subseafloor basalt occurs principally during the first 10 m.y. after basalt formation (Bach and Edwards, 2003). During this interval, the Fe(III)/ΣFe ratio of the bulk basalt in the database increases from ~0.15 to 0.35 and most of the sulfur is oxidized (Bach and Edwards, 2003). Whether this alteration is microbially mediated or not, it changes the redox habitability of the basalt. Alteration patterns are heterogeneous within and between cores, with greatest alteration in permeable zones, such as brecciated pillows (Bach and Edwards, 2003).

Recent studies of altered basalt at ODP Site 801 (~170 Ma) showed that alteration characteristics of this ancient oceanic crust are generally similar to those observed in much younger crust (e.g., ODP Hole 504B at ~6 Ma), suggesting that most alteration takes place when oceanic crust is young (Alt et al., 1992). If oxidative alteration ceases in 10–15 m.y. after crust formation, then the oceanic crust inhabitable by mineral-oxidizing microbes is limited to crust younger than 10–15 Ma.

Other evidence suggests that oxidative alteration need not be limited to crust younger than 10–15 Ma. Older crust has the potential for continued redox alteration, and it probably occurs in some geochemical/hydrological regimes. Geophysical measures of matrix density suggest that about half of all intergrain-scale crustal alteration in the uppermost basalt occurs in crust older than 10–15 Ma (Jarrard et al., 2003). In the equatorial Pacific, dissolved O2 and NO3 are present in basement as old as 35 Ma (D'Hondt et al., 2004). As much as 65% of the Fe in the Bach and Edwards (2003) remains as Fe(II) in shallow oceanic basalt older than 10 Ma. Whether this Fe(II) continues to be oxidized at very slow rates where exposed to dissolved O2 and NO3 or is physically inaccessible to oxidation remains to be determined.

Sediment thickness provides one explanation for the average change in redox alteration at 10–15 Ma. Once sedimentary cover is thick enough and anoxic enough to seal basement from contact with oxidized seawater, redox disequilibria (habitability) disappears and oxidative alteration ceases. Drilling basement of different ages beneath the unusually thin sediment cover in this region of the Pacific presents a unique test of this hypothesis. If a critical sediment thickness is necessary to curtail oxidative alteration, then the young (13.5 Ma), moderate (33.5 Ma), and old (84–125 Ma) basement sites of the proposed drilling effort will exhibit more intense or differing styles of alteration relative to other fast-spreading sites of comparable age and greater sediment cover (e.g., Site 801, ~400 m of sediment at ~170 Ma; ODP Site 1256, ~200 m of sediment at 15 Ma; and Hole 504B, ~250 m of sediment at ~6 Ma). If this is the case, the basalt of the South Pacific Gyre are redox habitable for many tens of millions of years after basalt formation.

Evolution of crustal hydrology and chemical habitability

The timing and distribution of shallow crustal alteration is intriguingly linked to evidence of the hydrologic evolution of subseafloor basalt. Seismic velocity and macro-porosity data suggest that porosity and bulk permeability of subseafloor basalt decrease rapidly during the first 10–15 m.y. after basalt formation (Jarrard et al., 2003). These decreases are thought to derive from secondary mineralization in the basalt. Mineralization reduces surface area within the crust and may limit electron donor and nutrient availability in the basaltic aquifer.

Global compilations of heat flow data indicate that advective heat loss is high in young seafloor and decreases with increasing age until ~65 Ma, where, on average, advective heat loss ceases (Stein and Stein, 1994) (Fig. F6). Fisher and Becker (2000) explain these observations by invoking closure of small-scale permeability in basement within 10–15 m.y. but ongoing flow for 65 m.y. in large fractures and faults that fill a relatively small volume of the rock. Three mechanisms are thought responsible for limiting hydrothermal circulation in old crust: (1) buildup of low-permeability sediment cover, which reduces basement relief and isolates the oceanic crust from overlying seawater; (2) ongoing mineralization, which decreases basement permeability with age; and (3) decreasing basal heat flux with age, which reduces the driving force for buoyant fluid flow. Basement permeability is commonly thought to ultimately control the cessation of hydrothermal circulation (Stein and Stein, 1994; Jarrard et al., 2003).

However, a growing body of evidence suggests that oceanic crust remains permeable enough to sustain advective heat loss through its life. This evidence includes the significant variance in age-dependent heat flow averages (Stein and Stein, 1994) (Fig. F6), large variations in heat flow survey data of Cretaceous-aged crust (Von Herzen, 2004; Fisher and Von Herzen, 2005) velocity logs, macroporosity data, matrix data (Jarrard et al., 2003), present-day fluid flow within ~132 Ma basement at ODP Site 1149 (Shipboard Scientific Party, 2000), and celadonite precipitation ages that indicate low-temperature fluid circulation at the Trodos ophiolite (Gallahan and Duncan, 1994). Given this evidence, the termination of the average heat flow deficit at ~65 Ma probably signifies that much of the open circulation between crust and ocean has largely stopped by then, rather than that hydrothermal flow has ended (Anderson and Skilbeck, 1981; Jacobson, 1992; Stein and Stein, 1994).

We hypothesize that advective fluid flow remains relatively vigorous even on old seafloor. If this is true, the apparent waning of hydrothermal circulation at ~65 Ma is controlled by sediment, rather than by basement permeability. The South Pacific Gyre offers a unique opportunity to test this hypothesis because its sediment is thin and discontinuous and its basement relief is relatively variable. If the hypothesis is true, large-scale permeability will be high regardless of basement age and heat flow data at the proposed sites will tend to deviate significantly from conductive values. Surface heat flow data from our recent survey (Fig. F6) and from scattered older measurements within 200 km of proposed Sites SPG-1A and SPG-4A suggest that an active flow system is present in the basement throughout the region of proposed drilling.

If this hypothesis is correct, the supply of the dissolved electron acceptors O2 and NO3 to the upper basement of the South Pacific Gyre may remain high long past the first 10–15 m.y. after basalt formation. If drilling of the lowermost sediment and the basalt shows the occurrence of these electron acceptors in the upper basement at our proposed sites, either (1) the metabolic habitability of South Pacific Gyre basalt remains sufficient to support life for as long as 100 m.y. or (2) the metabolic habitability of this basalt is ultimately controlled by the inaccessibility of electron donors in the basalt rather than by access to electron acceptors. If the first case applies, increased oxidative alteration will be evident in crust of increasingly greater age (e.g., proposed Site SPG-1A alteration will be greater than proposed Site SPG-4A alteration, which will be greater than proposed Site SPG-6A alteration). If the second case applies, oxidative alteration will be similar at all three sites.

Basement community composition

IODP drilling of the basalt at sites will provide a direct opportunity to test the existence and composition of microbial communities in oceanic crust of three very different ages (13.5, 33.5, and 80–125 Ma). Molecular and microscopic analysis of uncontaminated basalt from the site with youngest crust will test whether or not microbes take full advantage of the redox habitability of relatively young crust in open exchange with the overlying ocean. Molecular and microscopic analysis of uncontaminated basalt from the sites with older crust will test whether or not ongoing oxidation of the subseafloor basalt (by occurrence of reduced Fe and supply of dissolved O2 and NO3 by ongoing advection and diffusion through the sediment) sustains similar (or different) communities, albeit at slower rates of redox activity. Collectively, the study of these sites will track the compositional and functional evolution of basement communities >100 m.y. in a thinly sedimented region of fast-spreading crust, where we hypothesize supply of dissolved oxidants to continue for as long as 125 m.y. (or more).

Explanation of secondary objectives

Oceanic crust inputs to the subduction factory

The altered oceanic crust at our oldest basement sites (proposed Sites SPG-1A and SPG-2A) is outboard of the Tonga-Kermadec subduction zone. Characterizing the style and extent of alteration in this subducting plate will provide an important reference for assessing the role of altered oceanic crust in the subduction process. Of global arcs, the Tonga arc has perhaps the smallest sediment flux (Plank and Langmuir, 1998) and the altered oceanic crust in this system is proportionally more influential in the subduction process than at most other arcs. If the low sedimentation rate at this site has resulted in unusual alteration characteristics in the subducting oceanic crust, these differences may translate into geochemically distinct signatures in the magmas produced at the Tonga arc.

Regional tectonics

Proposed Sites SPG-1A and SPG-2A are centrally located in ocean crust accreted during the Cretaceous Normal Superchron (CNS). The tectonic history of this Australia-sized area is poorly constrained because correlatable magnetic seafloor anomalies are not present. Although limited bathymetry data suggest a general north–south spreading direction, the actual direction(s) and the presence of a failed rift within the region are poorly constrained. Drilling and logging ocean crust at either of these sites would address these questions. Radiometric dating of plagioclase within the recovered basalt would provide important age constraints (Koppers et al., 2003) and structural analysis of dipping flow units with Formation MicroScanner (FMS) logging data would provide a spreading direction (Pockalny and Larson, 2003). Resolving the tectonic history of this region is critical to understanding the effect of large igneous provinces (LIPs) on tectonic processes and whether the Ontong Java, Manihiki, and Hikurangi plateaus were created by one or multiple mantle plumes (Taylor, 2006).

Geodynamo

The causal mechanisms for the Cocos-Nazca spreading center are still debated. Some authors have proposed the presence of a strong magnetic field during superchrons (Larson and Olson, 1991; Tarduno et al., 2001), which would argue for an efficient geodynamo (i.e., large intensities during low reversal frequency). Others argue for a weak field (Loper and McCartney, 1986; Pick and Tauxe, 1993), which suggests increased convective vigor in the core would increase the reversal rate by generating frequent instabilities. If the latter model applies, the Cocos-Nazca spreading center record is low in paleointensity but may contain frequent reversals. To add further controversy, the limited number of paleointensity measurements from the Cocos-Nazca spreading center yield very different results for different methods (Pick and Tauxe, 1993; Tarduno et al., 2001). Drilling basement at proposed Site SPG-1A will provide important data and samples to test these models and methods. The paleointensity methods could be compared and the measurements would provide important data (in conjunction with the radiometric age) for the origin of the Cocos-Nazca spreading center and its relationship to the geodynamo.

Paleoceanography

Interstitial waters of the South Pacific Gyre represent a unique archive of glacial-aged water where relict NO3 can be used to test hypotheses of glacial–interglacial ocean-climate change with significantly lower uncertainty than through proxy measurements.

It has been hypothesized that changes in strength of the biological carbon pump caused the dominant variation in Earth's climate and atmospheric CO2 over the last 1 m.y. (Broecker, 1982). We will test the two principal models of this variation by reconstructing the preformed NO3 content and deep ocean δ15NO3 of the last glacial maximum (LGM) through measurement of O2, NO3, and δ15NO3 in the interstitial fluid of gyre sediment. Our tests utilize the fact that pore fluids from depths greater than ~30–50 mbsf (e.g., at proposed Sites SPG-1A and SPG-11B) are samples of paleobottom water that have effectively been out of diffusive contact with the ocean since the LGM.