• Subduction;
  • flux;
  • carbon dioxide;
  • potassium;
  • oceanic crust;
  • Ocean Drilling Program


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] The alteration of upper oceanic crust entails growth of hydrous minerals and loss of macroporosity, with associated large-scale fluxes of H2O, CO2, Cl, and K2O between seawater and crust. This age-dependent alteration can be quantified by combining a conceptual alteration model with observed age-dependent changes in crustal geophysical properties at DSDP/ODP sites, permitting estimation of crustal concentrations of H2O, CO2, Cl, and K2O, given crustal age. Surprisingly, low-temperature alteration causes no net change in total water; pore water loss is nearly identical to bound water gain. Net change in total crustal K2O is also smaller than expected; the obvious low-temperature enrichment is partly offset by earlier high-temperature depletion, and most crustal K2O is primary rather than secondary. I calculate crustal concentrations of H2O, CO2, Cl, and K2O for 41 modern subduction zones, thereby determining their modern mass fluxes both for individual subduction zones and globally. This data set is complemented by published flux determinations for subducting sediments at 26 of these subduction zones. Global mass fluxes among oceans, oceanic crust, continental crust, and mantle are calculated for H2O, Cl, and K2O. Except for the present major imbalance between sedimentation and sediment subduction, most fluxes appear to be at or near steady state. I estimate that half to two thirds of subducted crustal water is later refluxed at the prism toe; most of the remaining water escapes at subarc depths, triggering partial melting. The flux of subducted volatiles, however, does not appear to correlate with either rate of arc magma generation or magnitude of interplate earthquakes.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] Long-term average production of oceanic crust at spreading centers is balanced by subduction. The same is not true, however, for individual elements. Oceanic crustal alteration, associated with axial and ridge-flank hydrothermal circulation, modifies crustal abundances of several elements. For upper crust, the largest changes occur for structural water, carbon dioxide, and potassium [Staudigel et al., 1996], all of which are enriched by alteration.

[3] Subduction recycling of H2O, CO2, Cl, and K2O affects the geochemical evolution of the mantle, oceans, and atmosphere, impacting our understanding of a wide suite of global geochemical cycles [Peacock, 1990]. Volatile outgassing from MORB and arc magmas is probably the main source of the water and chlorine for early growth of the oceans [Rubey, 1951], but these fluxes are currently nearly balanced by loss to subduction. For no-growth continents, long-term sediment subduction should approximately equal arc magmatism [Armstrong, 1968, 1991]. Slab dehydration generates overpressures and reflux at shallow depths [Moore and Vrolijk, 1992] and partial melting at ∼100 km responsible for arc magmatism [Gill, 1981]. Subduction permanently removes CO2 that was temporarily isolated from the ocean/atmosphere system by deposition of carbonates and organic matter. Subduction recycling of volatiles and potassium are also likely to generate mantle heterogeneity and may affect compositions of arc magmas [Plank and Langmuir, 1993; Staudigel et al., 1996], though these effects are yet to be explored comprehensively.

[4] The subduction-related fluxes of this study contribute to the Geochemical Earth Reference Model (GERM), an interdisciplinary, long-term initiative with the much wider aims of characterizing the Earth's major chemical reservoirs and the fluxes between them [Staudigel et al., 1998a]. Subduction fluxes for all major elements have been estimated for both upper crustal extrusives [Staudigel et al., 1996] and overlying sediments [Plank and Langmuir, 1998]. This study incorporates geophysical models for age-dependent crustal alteration into calculations of fluxes for all subduction zones.

2. Age Dependence of Upper Crustal Properties

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[5] The evolution of hydrothermal circulation is well established qualitatively. At the ridge axis, magmatic heat drives vigorous and deep axial circulation [Humphris, 1995; Fisher, 1998]. Early off-axis hydrothermal circulation may be largely low-temperature (<30°C) with seafloor communication or higher-temperature (30–60°C) if sedimentation isolates the circulation from seawater [Fehn and Cathles, 1986; Davis et al., 1992; Humphris, 1995]. Hydrothermal circulation continues within ridge flanks for at least tens of millions of years (e.g., Anderson et al. [1977], Fisher et al. [1990], and many others). Off-axis waning of hydrothermal circulation is caused by decrease in crustal heat, accumulation of a relatively low-permeability sediment cover, and decrease in crustal permeability caused by crustal alteration [Anderson et al., 1977; Stein and Stein, 1994]. Of these three mechanisms, crustal alteration has received the most attention, perhaps because only it can account for the age-dependence of upper crustal seismic velocities [Schreiber and Fox, 1976, 1977]. My focus is on this crustal alteration.

[6] According to this conceptual model, alteration of oceanic crust begins as soon as crust is created. The combination of axial heat and highly fractured, permeable basalt promotes vigorous hydrothermal circulation, which causes crustal alteration. Alteration minerals eventually fill macroporosity (cracks and interpillow voids), thereby reducing porosity and permeability. Nevertheless, circulation on flanks is responsible for ∼70% of the advective heat loss from oceanic crust [Stein and Stein, 1994] and substantial geochemical exchange with seawater [Mottl and Wheat, 1994; Elderfield and Schultz, 1996].

[7] Low-temperature alteration patterns of upper oceanic crust have been determined from petrographic studies, core descriptions, and geochemistry; Honnorez [1981], Thompson [1991], Alt [1995, 1999], and Bach et al. [2003] provide comprehensive reviews. Early formation of celadonite and smectites (particularly saponite), often as intergrowths, is followed by precipitation of carbonates (mainly calcite). These three minerals are the dominant alteration minerals in upper oceanic crust; others include Fe-oxyhydroxides and zeolites. Early low-temperature alteration is oxidative, with high water-rock ratios and open seawater circulation. Reducing conditions generally follow, associated with decreases in both flow rates and exchange between seawater and pores.

[8] Basalt coring and logging by DSDP and ODP provide abundant information on physical properties of upper oceanic crust, despite generally low basement core recoveries and rare basement logging [Goldberg, 1997]. Most of the analyses of this section draw heavily from a synthesis of DSDP/ODP petrophysical data on aging of upper oceanic crust by Jarrard et al. [2003]. Of ∼1200 sites drilled by DSDP and ODP, 13 fulfill the following two criteria: (1) basement composed of normal oceanic crust formed in an open-ocean environment, and (2) velocity logging of at least 40 m of basement (Figure 1). Jarrard et al. [2003] confined their analyses to the top 300 m of basement, because most sites have only 90–200 m of velocity log and only four holes (395A, 418A, 504B, and 801C) penetrated more than 300 m of basement. A second search of the DSDP and ODP databases and publications (Initial Reports and Scientific Results) focused on index properties of basalts. Only sites on normal oceanic crust were considered, and data from multiple holes at the same site were combined. Useful index data are available for 25 sites (Figure 1). These analyses of crustal alteration patterns exclude data from ophiolites, for two reasons: (1) most are formed in supra-subduction settings that are nonrepresentative of normal, open-ocean crust, and (2) ophiolitic alteration is generally more intense, with much higher total fluid flux, than normal crustal alteration [Alt and Teagle, 2000]. Sedimented ridges and back-arc spreading are also excluded, because their hydrothermal circulation histories probably differ from that of open-ocean crust.


Figure 1. Locations of DSDP and ODP sites with significant penetrations into normal oceanic crust and with either downhole logging or core physical properties. Also shown are locations of Holes 504B (Figure 3), Site 801 (Figure 4), and gabbro Hole 735B.

Download figure to PowerPoint

[9] Basalt physical properties at intergranular core-plug scale may differ from those at the meter scale measured by logs, because of differences in type of porosity and in alteration history. At plug scale, porosity consists mainly of vesicles and microcracks, whereas log-scale porosity also includes fractures and interpillow voids. At plug scale, alteration converts glass, olivine, and plagioclase mainly to clay minerals, thereby decreasing matrix density and matrix velocity. Porosity may increase due to this hydration, or decrease due to precipitation. At log scale and larger, porosity is thought to decrease due to filling of cracks and interpillow voids by alteration minerals. Consequently, a comparison of plug-scale and log-scale basalt physical properties may yield insights concerning the scales and varieties of alteration.

2.1. Permeability, Porosity, and Alteration

[10] Permeability is the architect of the crust's hydrothermal systems. Nearly all measurements of the permeability of upper oceanic crust are averages for portions of a borehole tens to hundreds of meters in extent. Fisher [1998] provided an excellent review and synthesis of these bulk permeability measurements and their implications. This large-scale crustal permeability is a critical control on hydrothermal circulation patterns and associated heat flux, because the volumetrically dominant flux is channelized within the most permeable zones, particularly large open fractures [Fisher, 1998; Fisher and Becker, 2000] and breccias [Bach et al., 2003]. Bulk permeabilities for the top 1200 m of oceanic crust suggest two main layers: the top ∼500 m has permeabilities of ∼10−14–10−13 m2, and the next 700 m has permeabilities of ∼10−17 m2 [Fisher, 1998]. The intense intergranular alteration of pillows and flow margins, despite intergranular permeabilities that are more than four orders of magnitude lower than the fracture and interpillow permeabilities measured by packers, occurs via both microbial activity [e.g., Furnes and Staudigel, 1999; Furnes et al., 2001; Staudigel et al., 1998b] and diffusion. The lower limit for sufficient diffusion to produce significant alteration may be ∼10−19 m2, based on the observation that lower intergranular permeabilities are encountered mostly in fresh, massive basalts [Karato, 1983a, 1983b; Johnson, 1979a; Hamano, 1979; Jarrard et al., 2003].

[11] ODP Hole 801C, which obtained the world's oldest section of in situ, normal oceanic crust, provides the opportunity to examine relationships among hydrologic properties (porosity, permeability, fluid flow), crustal alteration, and geophysical properties, at both core-plug and downhole-log scales [Busch et al., 1992; Jarrard et al., 1995, 2003]. Within these upper crustal basalts, higher porosity is responsible for higher permeability and therefore higher fluid-flow rates. High fluid flux, in turn, fosters alteration, particularly the hydration reactions that generate clays. Consequently, porosity is well correlated both with hydration intensity determined by light-absorption spectroscopy (LAS) (significant at 99% confidence level) and with matrix density (significant at 99% c.l.). Because hydration substantially reduces matrix density, matrix density is correlated with LAS-based hydration (significant at 99% c.l.) [Jarrard et al., 2003].

2.2. Timing of Waning of Hydrothermal Circulation: Previous Evidence

[12] Geophysical techniques disagree on the temporal evolution of hydrothermal circulation and resulting alteration in oceanic crust. Cementation associated with ridge-flank hydrothermal circulation appears to terminate as early as ∼5–10 Ma, based on global data syntheses of upper crustal seismic velocities [Grevemeyer and Weigel, 1996; Carlson, 1998]. Similarly, crustal permeabilities decrease exponentially from 1 Ma to 8 Ma; subsequent changes (if any) are undetermined [Fisher and Becker, 2000].

[13] In contrast, the global heat flow synthesis of Stein and Stein [1994] indicated that the heat flow deficit (theoretical minus observed) gradually decreases with age, disappearing at about 65 Ma. The deficit indicates advection via open circulation between seawater and the crust; beyond 65 Ma, heat flow is dominantly conductive, with an age-dependent decrease in heat flow variance that suggests continued fluid flow [Stein and Stein, 1994]. Von Herzen [2003] reanalyzed 58 detailed heat flow surveys from old crust and concluded that ongoing hydrothermal circulation is common in Mesozoic crust; it is likely in half of the surveys on crust 65–95 Ma, and in a third of those on >95-Ma crust. Heat flow data cannot unambiguously determine whether hydrothermal circulation in old crust is closed-cell (no interchange with seawater) or slow open-cell [Fisher, 1998], yet the two differ crucially in geochemical mass balance. For closed-cell fluid flow, precipitation equals solution, and crustal geophysical properties may be unchanged by ongoing alteration.

[14] These conflicting indications of when off-axis hydrothermal circulation wanes have been partly reconciled. Fisher and Becker [2000] suggested that the rapid initial increase in velocity and decrease in permeability are not necessarily incompatible with heat flow evidence for persistence of open-cell convection. They hypothesized that hydrothermal circulation continues to 65 Ma, in localized channels within the crust that are too small a percentage of total upper crust for significant effects on average crustal velocity. Crustal alteration beyond ∼5–10 Ma is too subtle for most geophysical methods to resolve, but the next three sections examine continuing geophysical changes that are detectable.

2.3. Core Alteration and Matrix Density Versus Age

[15] Johnson and Semyan [1994] examined the possibility of age-dependent changes in velocity, porosity, density, oxidation state, and hydration, based on a compilation of core-plug measurements from DSDP and ODP basalts. The increase of crustal alteration with increasing age is most evident among parameters sensitive to hydration. Age-dependent oxidation is clearest for 0.1–5 Ma and is not resolved beyond ∼10–30 Ma [Johnson and Semyan, 1994; Zhou et al., 2001b]. Analyses of bound water, or H2O+, provide a semiquantitative measure of the extent of alteration-induced hydration [Alt et al., 1992]. The H2O+ compilation of Johnson and Semyan [1994] confirmed earlier indications [Hart, 1970, 1973; Donnelly et al., 1979a; Muehlenbachs, 1979] that hydration increases with increasing age (R = 0.35; significant at 95% confidence level), but dispersion is too high to identify when most of this hydration occurs.

[16] Matrix density is the average density of the minerals forming the solid part of the rock, including any alteration minerals. Matrix densities provide the strongest demonstration of systematic increase in alteration versus time (Figure 2b) (R = −0.70; significant at 99.9% confidence level), indicating that approximately half of all intergranular-scale crustal alteration occurs after the first 10–15 Ma. Site 597, where sampling concentrated on a massive flow [Shipboard Scientific Party, 1985] and matrix density is anomalously high, is excluded. The simplest interpretation of this pattern is that matrix density continues to decrease, at a rate proportional to logarithm of age, throughout the period 0–167 Ma. However, data dispersion in the time period 10–70 Ma is high, and the hypothesis of no change beyond about 40–60 Ma cannot be rejected.


Figure 2. Age-dependent geophysical properties of upper oceanic crust. Each point represents one DSDP or ODP site. Regression lines are used to predict crustal properties at subduction zones, based on crustal age. A–C are modified from Jarrard et al. [2003].

Download figure to PowerPoint

2.4. Velocity Versus Age

[17] Compilations of published seismic experiments from throughout the world ocean demonstrated that seismic velocities of the upper oceanic crust increase rapidly for the first 5–8 Ma, from an average of 2.3 km/s at the spreading center to 4.3–4.4 km/s at 5–10 Ma (Figure 2a) [Grevemeyer and Weigel, 1996; Carlson, 1998]. The amount of associated crustal alteration is undetermined. Velocity modeling suggests that the large near-ridge velocity increases may be accomplished by changing the shape of pore spaces via secondary mineralization with relatively small overall porosity reduction [Wilkens et al., 1991; Shaw, 1994; Moos and Marion, 1994]. Beyond 10 Ma, no significant change in upper crustal velocity is resolvable with seismic velocity data [Carlson, 1998]; data dispersion is too high to resolve systematic changes of <1 km/s. DSDP and ODP velocity logs, with an age range of 6–167 Ma that is beyond any seismically resolvable age dependence, indicate that crustal velocity increase continues (Figure 2a, significant at 95% c.l.), rising from 4.2–4.7 km/s at 6 Ma to 5.0–5.1 km/s for >100 Ma. This gradual increase in large-scale velocity occurs despite systematic velocity decrease at the intergranular scale, as indicated by core-plug velocities for 0–100 Ma [Christensen and Salisbury, 1973; Johnson and Semyan, 1994; Jarrard et al., 2003]. Precipitation in cracks and interpillow voids must be dominant over intergranular-scale alteration in controlling the evolution of large-scale crustal velocity.

2.5. Macroporosity Versus Age

[18] Plugs reveal intergranular patterns, but they tend to miss cracks and crack-filling that potentially are major controls on large-scale velocity [Hyndman and Drury, 1976; Anderson and Zoback, 1982]. Evaluation of crack porosity based on incomplete core recovery is difficult [Johnson, 1979b]. If much of basalt porosity is in the form of large-scale voids and fractures that are not sampled by 10-cm3 core plugs, then comparison of core and log measurements of porosity has the potential of detecting this macroporosity. I use core and log velocity for macroporosity determination, because velocity logs provide the most robust, unbiased measure of in situ porosities of oceanic basalts.

[19] Before comparing atmospheric-pressure core-plug velocities to in situ log velocities, pressure effects must be considered. Rebound, the core expansion that accompanies change from in situ to laboratory pressure, can reduce plug velocities [e.g., Nur, 1971; Bourbié et al., 1987]. Rebound is generally minor in oceanic extrusive basalts, because burial depths are much too small to induce microcracks. The extensive suite of core velocity measurements for Hole 504B [Wilkens et al., 1983; Christensen and Salisbury, 1985; Christensen et al., 1989; Iturrino et al., 1995; Salisbury et al., 1996] shows that the average pressure-induced percentage difference between atmospheric and in situ velocities for the upper 600 m of basement is only −0.3% based on differential pressures, −0.6% based on lithostatic pressures, and −2.0% based on confining pressures. In contrast, the two diverge at greater depths: at about 1100 m subbasement, ∼300 m into the dikes, the unmistakable signature of microcracks appears and atmospheric-pressure measurements are no longer a useful indicator of in situ velocities (Figure 3). Consequently, I assume that atmospheric-pressure velocities can be used for macroporosity calculations within this and other extrusive sections, whereas high-pressure measurements would be required for dikes and gabbros.


Figure 3. Downhole K2O variations, seismic layers, lithologic units, and P wave velocity rebound at Hole 504B. Sources in text. K2O variations are shown with red dots for core XRF analyses, black lines for well log, and vertical green line for assumed pre-alteration K2O concentration. Velocity rebound is shown as percentage ratio of velocity at atmospheric pressure compared to that at in situ (lithostatic) pressure.

Download figure to PowerPoint

[20] Figure 2c compares core-plug and log velocities for the 13 sites with significant basement penetration. Because a systematic difference between the two is thought to result primarily from the influence of macroporosity on logs but not on cores, I used the Wyllie et al. [1956] time-average equation to express the difference between each core/log velocity pair as an apparent macroporosity. Matrix velocity may also decrease with age [Jarrard et al., 2003] but no macroporosity bias results, because both core-plug and log velocities are similarly affected. Total macroporosities exhibit a decrease with increasing age (R = −0.87) that is significant at the 99.9% confidence level, and this decrease appears to continue throughout the available age range (Figure 2c). Initial macroporosities of 8–10% eventually are reduced to near zero. This systematic macroporosity decrease contrasts with the absence of significant age-dependent porosity change among core plugs [Johnson and Semyan, 1994; Jarrard et al., 2003]. In contrast to the precipitation in cracks and interpillow voids that causes macroporosity decrease, intergranular alteration can either fill small cracks and vesicles or generate porosity by breakdown of individual minerals such as olivine.

2.6. CO2 Versus Age

[21] Ridge volcanism releases CO2 at axial hydrothermal vents, mainly via degassing at magma chambers rather than via crustal alteration [Gerlach, 1989]. Residual CO2 concentrations within the basalts are minor, only ∼0.045 wt.% [Gerlach, 1989]. Subsequent low-temperature alteration adds CO2 to the upper crust, primarily in the form of CaCO3, using CO2 from seawater and CaO mostly from the basalts [Staudigel et al., 1989, 1996]. This carbon sink has been demonstrated to be an important term in global carbon budgets [Staudigel et al., 1989; Peacock, 1990; Alt and Teagle, 1999].

[22] Progressive CO2 enrichment of oceanic crust can be quantified by analysis of DSDP and ODP sites, but avoidance of sampling bias is non-trivial. Most whole-rock CO2 analyses target the least weathered samples and avoid carbonate-filled cracks. Core recovery is also biased, losing some of the larger-scale crack filling. Two studies of crustal CO2 have succeeded in minimizing these biases: Staudigel et al. [1996] and Alt and Teagle [1999].

[23] Adjacent Holes 417A, 417D, and 418A in the west-central Atlantic provide an unusually good record of the alteration of old oceanic crust, due to penetration of almost the entire extrusive section, relatively high core recovery (∼70%), and sampling of both hydrothermal downflow and upflow zones. Staudigel et al. [1996] took several steps to obtain representative samples of downhole geochemical variations within these holes: they sampled a range of alteration states for all extrusive styles (flows, pillows, and volcaniclastics), including vein materials, and then they blended the results using best estimates of the in situ proportions of alteration states and extrusive styles. These composite CO2 concentrations provide an excellent demonstration of the difference between average geochemical analyses and representative samples; the former are almost all 0–2 wt.% CO2, whereas the latter exhibit a systematic decrease downhole: 5.6% for 0–100 m subbasement, 5.0% for ∼130–290 m, and 1.0% for ∼455–540 m. The resulting weighted average for the upper extrusive section (0–300 m subbasement) is 5.2%, similar to an earlier 5.0% value by Staudigel et al. [1989], and the overall average CO2 concentration of the extrusives at these three holes is 2.95% [Staudigel et al., 1996].

[24] Alt and Teagle [1999] combined best estimates of bulk-rock and vein CO2 to obtain representative samples of the average CO2 contents of upper oceanic crust at Sites 504, 896, 843, and 801. Bulk-rock CO2 measurements were averaged for different alteration types, then these results were weighted based on observed proportions of the alteration types, producing an average bulk-rock CO2 for each crustal section. The amount of vein carbonate was calculated from detailed core descriptions of carbonate vein abundances and thicknesses plus breccia carbonate abundance. These results demonstrated that carbonate precipitation in veins continues beyond the 6 Ma age of Sites 504 and 896, resulting in enrichment of total CO2 content of the upper extrusive section.

[25] Figure 2d combines the upper crustal CO2 determinations of Alt and Teagle [1999] with that of Staudigel et al. [1996]. The pattern of progressive CO2 enrichment extends far beyond 10 Ma, once hypothesized to be the conclusion of calcite precipitation [Staudigel et al., 1981]. Alt and Teagle [1999] emphasized that these data establish an increase in CO2 between 6 Ma and 110 Ma but not the specific timing of this increase. Given the evidence of continuing upper-crustal alteration based on velocity (Figure 2a), macroporosity (Figure 2c), matrix density (Figure 2b), and variance of global heat flow [Stein and Stein, 1994], however, the simplest explanation of age-dependent CO2 is that CO2 enrichment is ongoing but waning, with a logarithm of time dependence (Figure 2d; R = 0.88).

[26] Cretaceous atmospheric CO2 levels were an order of magnitude higher than Neogene levels [Ekart et al., 1999; Pearson and Palmer, 2000; Berner and Kothavala, 2001], possibly explaining the pattern of Figure 2d without requiring continued off-axis carbonate precipitation. Within-site correlations of highest whole-rock and vein CO2 with highest present-day porosity and presumably permeability are consistent with – but cannot prove – ongoing, progressive fluid flow and associated carbonate enrichment. Petrographic studies indicate that carbonate precipitation generally postdates smectite alteration [Alt and Teagle, 1999]. This observation is consistent with the regressions of Figure 2, which suggest that half of all matrix density lowering (smectite/celadonite alteration) occurs after the first 15 m.y., whereas half of all CO2 addition occurs after the first 25–30 m.y. The calculated modern CO2 subduction fluxes of this study depend only on the pattern of Figure 2D, not its mechanism, whereas modern precipitation flux is sensitive to both.

2.7. Potassium Versus Age

[27] Low-temperature crustal alteration processes include oxidation, hydration, and alkali fixation. The latter two generate clay minerals, the most abundant alteration products [e.g., Donnelly et al., 1979a; Andrews, 1980; Gillis and Robinson, 1988; Alt and Honnorez, 1984; Alt et al., 1986]. Saponite (a smectite clay) and celadonite (a mica) are commonly formed during low-temperature diagenesis [e.g., Donnelly et al., 1979b; Alt and Honnorez, 1984], usually incorporating potassium extracted from seawater [Hart, 1969; Andrews, 1980]; an exception is saponite formed by alteration of interstitial glass [Zhou et al., 2001a]. Consequently, altered basalts are generally higher in potassium than unaltered basalts [Hart, 1969; Hart et al., 1974]. Deep penetration of old oceanic crust at Holes 417A/417D/418A provided the first compelling evidence of widespread alteration-induced K2O enrichment in the upper crust [Donnelly et al., 1979a, 1979b].

[28] Potassium enrichment is greatest in the most porous, permeable zones. Figure 4, which overlays K2O and velocity logs versus depth for the Hole 801C tholeiites, demonstrates that potassium variations are closely linked to porosity changes (velocity is used here as a porosity indicator). The spectral gamma-ray log of potassium is here converted to K2O weight percent by multiplying by 1.2. The lowest observed K2O contents, ∼0.07–0.12%, are comparable to values typical of unaltered MORB tholeiites, suggesting that the most massive, low-porosity units are minimally altered. K2O maxima of ∼1% result from a high proportion of potassium-bearing alteration minerals and occasional sediment interlayers. This strong inverse correlation between velocity and potassium-bearing alteration minerals is not inconsistent with the earlier observation that log velocity increases with time and crustal alteration (Figure 2a): the highest-porosity zones experience the most alteration and associated porosity reduction, yet they remain more porous than fresher, more massive intervals.


Figure 4. Comparison of velocity and potassium logs for Hole 801C tholeiites. The strong inverse correlation between these logs indicates that precipitation of potassium-rich clays and celadonite occurs mainly in the most porous (lowest-velocity) zones. Modified from Jarrard et al. [2003].

Download figure to PowerPoint

[29] Only seven of the thirteen logged basement sites in normal oceanic crust have potassium logs. These logs show no consistent pattern of potassium increase or decrease with depth. Among these seven sites, average log-based potassium content is strongly correlated with average core-based matrix density (R = −0.86; significant at 99% confidence level) (Figure 5). This correlation, like the Hole 801C correlation between potassium and velocity logs (Figure 4), demonstrates that alteration entails potassium enrichment. Also shown on Figure 5 are results for Site 768, which may be the most altered thick sequence of oceanic crust yet sampled by DSDP or ODP [Shipboard Scientific Party, 1990a]. This crust was formed by back-arc spreading and is consequently excluded from these analyses of normal crustal alteration, but it illustrates extreme alteration and associated potassium enrichment.


Figure 5. Black dots: average core-based matrix density versus average log-based K2O concentration, for upper oceanic crust at all DSDP/ODP sites with both data types; regression line is also shown. Dashed lines and associated labels identify 3-component (fresh basalt, saponite, celadonite) mixing models. For prediction of K2O and structural water at subduction zones, the best fit alteration mineral assemblage of 80% saponite and 20% celadonite is assumed.

Download figure to PowerPoint

[30] Despite the excellent correlation of Figure 5, the temporal evolution of potasium enrichment is an outstanding problem. Potassium-rich celadonite is generally one of the earliest-formed alteration minerals [Alt, 1999], and the few sites with potassium logs show only a fair correlation (R = 0.47) of potassium with logarithm of crustal age [Jarrard et al., 2003].

2.8. Alteration Mineralogy Model

[31] Geochemical changes associated with the progressive alteration of oceanic crust can be modeled as an age-dependent mix of fresh basalt and alteration minerals. I assume that the dominant alteration minerals in the upper extrusives are saponite and celadonite, though this is only a first-order approximation. Given the matrix densities and K2O contents of smectite, celadonite, and fresh basalt, one can compute the proportions of these three components from the relationship between matrix density and K2O (Figure 5).

[32] The average matrix density of fresh basalt is 3.01 g/cc, based on applying the regression of matrix density on age (Figure 2b) to ages <1 Ma. The local value depends on concentration of dense phenocrysts such as olivine and pyroxene. I assume a K2O content of 0.085 weight percent, based on analyses of fresh basalts from Holes 417A/417D/418A [Spivack and Staudigel, 1994; Staudigel et al., 1996] and 801C [Shipboard Scientific Party, 2000a]. Average unaltered N-MORB [Hart et al., 1999] and fresh glasses from dredged basalts [Jochum et al., 1983] are slightly higher: 0.106% and 0.12%, respectively. The freshest basalts from Holes 504B and 896A have an average K2O of 0.04% [Shipboard Scientific Party, 1992], much lower than typical MORB. This depletion may slightly affect the lowest two points on Figure 5, but the overall influence of magmatic K2O variations is clearly minor in comparison to alteration-induced variations.

[33] The matrix density of celadonite is assumed to be 2.56 g/cc, the same as that of glauconite [Serra, 1986], based on compositional similarity of the two minerals [Buckley et al., 1978]. Celadonite has an average K2O content of about 9.07%, based on 13 analyses by Buckley et al. [1978]. Saponite has an average K2O content of about 0.21%, based on 35 analyses of Honnorez et al. [1983].

[34] The structural interlayer water of smectites can vary from 10% to 25% by weight or higher [Ransom and Helgeson, 1995]. Furthermore, sample drying, required for either matrix-density measurement or chemical analysis, drives off not only pore water but also some of the loosely held structural interlayer water. Consequently, measured matrix densities are biased upward [Brown and Ransom, 1996] and H2O+ analyses are too low. Measured structural-water content of saponite averages 12 wt.%, based on 35 analyses by Honnorez et al. [1983], but I assume that the true value is 20 wt.%. Brown and Ransom [1996] estimated that the true water content of largely authigenic smectites offshore Barbados is 20 wt.%, and Schlumberger used neutron and density log analysis to propose that montmorillonites have a water content of 41% by volume [Serra, 1986], or 20 wt.%. Applying this estimate to basalt saponites, however, could be inaccurate by as much as 5 wt.%. The assumed bound-water content of 5% for celadonite is based on only one measurement by Buckley et al. [1978], because most analyzed oceanic crustal celadonites are a mix of celadonite and smectite [Andrews, 1980] (e.g., water contents of 9.5–11% for mixed celadonite/smectite analyses of Buckley et al. [1978]).

[35] For calculation of alteration mineral percentages using Figure 5, it is appropriate to use measured rather than true saponite matrix density, based on measurements with similar sample drying procedure to those applied to basalts: 24 hours heating at ∼110°C. Smectite sediments from Site 1222 have an average measured matrix density of 2.51 g/cc (std. dev. 0.05, N = 17) [Shipboard Scientific Party, 2002]. I assume that smectites within basalts are similar in water content to these and have an apparent matrix density of 2.50 g/cc.

[36] The dashed lines on Figure 5 show expected positions on a K2O versus matrix density plot of mixtures of fresh basalt, saponite, and celadonite. Observed data points, including Site 768, lie along a line for a mixture of fresh basalt and an alteration-mineral assemblage of 80% saponite and 20% celadonite; the data regression line lies virtually on top of that mixture line. Consequently, I assume that the alteration minerals in extrusive basalts are – to a first approximation −80% saponite and 20% celadonite. Other alteration minerals are present, but I assume that they either are similar in properties to these minerals or have a minor effect on total K2O and matrix density. For example, smectite is dominant in both early oxidizing and later reducing alteration, but early celadonite can be followed by zeolites with similar K2O and matrix density. This model composition for alteration minerals, along with their K2O contents and matrix densities, permits estimation of the K2O content of an extrusive sequence, based on the age-dependent function for matrix density (Figure 2b). The direct regression of K2O on matrix density (Figure 5) yields an almost identical result for predicted K2O and total global K2O flux. I use the mineralogic model instead because of its geologic grounding and because it predicts both K2O and structural water from matrix density (Table 1).

Table 1. Parameters for Flux Calculations
Trench lengthkmL meas.
Converg. rate, meas.
Crustal ageMaA meas.
Fluid densityg/cm3ρf 1.02
Zone thicknessmTssed. meas.
  Tuupper ext.300
  Tllower ext.300
Microporosity%ϕμ,uupper ext.7.8
  ϕμ,llower ext.5.1
Macroporosity%ϕm,uupper ext.ϕm,u = 13.01 − 5.625*log(A)
  ϕm,llower ext.ϕm,l = ϕm,u/2
Total porosity%ϕtot,uupper ext.ϕtot,u = ϕμ,u + ϕm,u
  ϕtot,llower ext.ϕtot,l = ϕμ,l + ϕm,l
  ϕtot,ddikesϕtot,d = ϕμ,d + ϕm,d
  ϕtot,ggabbrosϕtot,g = ϕμ,g + ϕm,g
Matrix densityg/cm3ρma,uupper ext.ρma,u = 3.01 − 0.0631*log(A)
  ρma,llower ext.ρma,l = 3.01 − 0.5*0.0631*log(A)
Bulk densityg/cm3ρb,ssed.meas.
  ρb,uupper ext.ρb,u = (0.01*ϕtot,u)*ρf + (1 − 0.01*ϕtot,u)*ρma,u
  ρb,llower ext.ρb,l = (0.01*ϕtot,l)*ρf + (1 − 0.01*ϕtot,l)*ρma,l
  ρb,ddikesρb,d = (0.01*ϕtot,d)*ρf + (1 − 0.01*ϕtot,d)*ρma,d
  ρb,ggabbrosρb,g = (0.01*ϕtot,g)*ρf + (1 − 0.01*ϕtot,g)*ρma,g
K2Owt. %K2Ossed. meas.
  K2Ouupper ext.K2Ou = 11.58 − 3.823*ρma,u
  K2Ollower ext.K2Ol = 11.58 − 3.823*ρma,l
CO2wt. %CO2ssed. meas.
  CO2uupper ext.CO2u = −1.55 + 2.493*log(A)
  CO2llower ext.CO2l = CO2u/2
Structural H2Owt. %H2OSssed. meas.
  H2OSuupper ext.H2OSu = (103.1 − 34.27*ρma,u) + (0.17*(13.01 − ϕm,u))
  H2OSllower ext.H2OSl = (103.1 − 34.27*ρma,l) + (0.17*(13.01 − ϕm,l))
pore H2Owt. %H2OPssed. meas.
  H2OPuupper ext.H2OPu = ϕtot,ufb,u
  H2OPllower ext.H2OPl = ϕtot,lfb,l
  H2OPddikesH2OPd = ϕtot,dfb,d
  H2OPggabbrosH2OPg = ϕtot,gfb,g
chloridewt. %CleachCl = 0.019*H2OP
K2O fluxg/yrFK2OeachFK2O = K2O*(1 − 0.01*H2OP)*T*ρb*L*V*104
CO2 fluxg/yrFCO2eachFCO2 = CO2*(1 − 0.01*H2OP) *T*ρb*L*V*104
structural H2O fluxg/yrFH2OSeachFH2OS = H2OS*(1 − 0.01*H2OP)*T*ρb*L*V*104
pore H2O fluxg/yrFH2OPeachFH2OP = H2OP*T*ρb*L*V*104
chloride fluxg/yrFCleachFCl = 0.019*FH2OP

[37] Both matrix density functions – versus age and versus K2O – are based on core-plug matrix density rather than log matrix density. However, large-scale matrix density may differ from intergranular matrix density because the amount or kind of fracture-fill minerals differs from intergranular alteration minerals. This difference does not bias the K2O value, which is log-based and therefore representative, but it does mean that computation of structural water is accurate only for intergranular alteration minerals. When macroporosity of old basalts has been filled by alteration minerals, large-scale matrix density is modified slightly from intergranular matrix density. Fortunately, the percentage of filled macroporosity is so small that its influence on total matrix density and structural-water estimates is minor; nevertheless, I employ a correction for this effect (Table 1).

3. Data Set: Global Subduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

3.1. Subduction Zones

[38] Nearly all of the world's currently active subduction zones are included in this analysis (Table 2), regardless of whether the subducting oceanic crust formed in an open-ocean or back-arc environment. Continent/continent collisions (e.g., India/Asia) and portions of a subduction zone undergoing continental collision (e.g., Aegean, E. Sunda) are excluded. With few exceptions, the subduction zones are those tabulated and analyzed by Jarrard [1986] and von Huene and Scholl [1991], though some names have been updated. The total length of active subduction zones in this study is 44,450 km, close to the 43,250 km total of von Huene and Scholl [1991]. Total ridge length is ∼49,000 km [Kominz, 1984]. The total length of convergent plate boundaries, including ∼10% continent-continent collisions, is also ∼49,000 km, so global average convergence and divergence rates are nearly equal.

Table 2. Subduction Zones Analyzed and Their Calculated Fluxesa
TrenchLength, kmRefConv. Rate, mm/yrRefAge, MaRefMw ThrustSubducting Sediment FluxesIgneous Crust Fluxes
K2O, g/yrCO2, g/yrH2O+, g/yrPore H2O, g/yrChloride, g/yrK2O, g/yrCO2, g/yrH2O+, g/yrPore H2O, g/yrChloride, g/yr
Aegean850l5b90l      7.00E + 102.05E + 111.11E + 125.43E + 111.03E + 10
Alaska1490a56a49a9.21.46E + 1205.00E + 123.80E + 137.21E + 111.31E + 123.35E + 122.11E + 131.12E + 132.13E + 11
Aleutian, E.1246a59a56a9.15.21E + 1101.98E + 121.75E + 133.33E + 111.16E + 123.09E + 121.87E + 139.79E + 121.86E + 11
Aleutian, W.1102a18a72a7.1     3.21E + 118.99E + 115.13E + 122.59E + 124.91E + 10
Andaman1089a23a,d83a5.72.93E + 127.17E + 101.24E + 133.62E + 136.89E + 114.10E + 111.18E + 126.54E + 123.22E + 126.13E + 10
Antilles, N.400e24e87a6.14.25E + 101.39E + 102.98E + 111.40E + 122.65E + 101.58E + 114.60E + 112.51E + 121.23E + 122.34E + 10
Antilles, S.400e24e87a6.16.19E + 111.77E + 102.05E + 129.48E + 121.80E + 111.58E + 114.60E + 112.51E + 121.23E + 122.34E + 10
C. America1506a72a16a7.82.60E + 111.02E + 131.76E + 123.63E + 136.91E + 111.54E + 122.77E + 122.54E + 131.60E + 133.03E + 11
Cascadia990a36a5a7.21.69E + 1206.46E + 123.02E + 135.73E + 114.53E + 113.86E + 117.73E + 125.71E + 121.09E + 11
Chile, C.1306a71a23a9.5     1.36E + 122.80E + 122.23E + 131.33E + 132.52E + 11
Chile, N.1579a74a41a8.5     1.80E + 124.42E + 122.92E + 131.60E + 133.03E + 11
Chile, S.1218a16a16a5.7     2.76E + 114.98E + 114.57E + 122.87E + 125.45E + 10
Colombia1355a65a21a8.82.96E + 105.08E + 121.58E + 121.62E + 133.08E + 111.28E + 122.56E + 122.10E + 131.27E + 132.41E + 11
Cotabato500k20f42n,i      1.54E + 113.81E + 112.50E + 121.36E + 122.59E + 10
Hikurangi794a21a100a7.7     2.77E + 118.30E + 114.40E + 122.11E + 124.01E + 10
Izu-Bonin1167a50g145a6.64.06E + 111.89E + 122.45E + 121.50E + 132.85E + 119.99E + 113.20E + 121.58E + 137.15E + 121.36E + 11
Japan, NE1061a76a132a8.24.79E + 1102.84E + 121.81E + 133.44E + 111.37E + 124.32E + 122.17E + 139.96E + 121.89E + 11
Java1200f71b100f7.86.29E + 112.69E + 112.06E + 121.70E + 133.23E + 111.42E + 124.24E + 122.25E + 131.08E + 132.05E + 11
Kamchatka900a74a115a9.01.84E + 1102.90E + 121.56E + 132.96E + 111.12E + 123.44E + 121.77E + 138.33E + 121.58E + 11
Kermadec1422a52a100a8.02.00E + 1107.74E + 111.20E + 132.28E + 111.23E + 123.68E + 121.95E + 139.37E + 121.78E + 11
Kurile1243a75a128a8.55.54E + 1103.29E + 122.09E + 133.98E + 111.58E + 124.95E + 122.50E + 131.16E + 132.19E + 11
Makran950c37b97d 6.79E + 121.23E + 131.28E + 136.10E + 131.16E + 125.82E + 111.73E + 129.26E + 124.46E + 128.48E + 10
Manila1050c10c30h      1.58E + 113.54E + 112.57E + 121.47E + 122.80E + 10
Mariana1812a47e134a7.87.90E + 111.52E + 123.94E + 121.73E + 133.30E + 111.45E + 124.58E + 122.29E + 131.05E + 132.00E + 11
Mexico1383a49a9a8.07.96E + 1006.47E + 119.31E + 121.77E + 119.11E + 111.23E + 121.53E + 131.04E + 131.98E + 11
Nankai824a38g23a8.16.25E + 1101.21E + 124.82E + 129.16E + 104.59E + 119.47E + 117.53E + 124.48E + 128.52E + 10
Negros400k20c16m      1.13E + 112.04E + 111.88E + 121.18E + 122.24E + 10
New Britain600f110c,o50d      1.03E + 122.67E + 121.67E + 138.87E + 121.69E + 11
Peru1599a65a37a8.17.00E + 101.35E + 125.09E + 111.05E + 132.00E + 111.59E + 123.79E + 122.58E + 131.43E + 132.72E + 11
Philippine1509a64g43a8.08.64E + 104.90E + 113.72E + 119.50E + 121.80E + 111.49E + 123.71E + 122.42E + 131.31E + 132.50E + 11
Ryukyu1153a57g46a6.32.28E + 1106.08E + 118.54E + 121.62E + 111.02E + 122.59E + 121.65E + 138.90E + 121.69E + 11
San Cristobal1050c49p50d      8.07E + 112.08E + 121.30E + 136.92E + 121.31E + 11
Scotia1005a44a57a7.01.30E + 1104.70E + 116.59E + 121.25E + 117.01E + 111.87E + 121.13E + 135.88E + 121.12E + 11
Sulawesi, N.600c30d42n,i      2.78E + 116.87E + 114.50E + 122.46E + 124.66E + 10
Sulu500k20f16m      1.42E + 112.55E + 112.35E + 121.47E + 122.80E + 10
Sumatra2462a50a61a7.16.22E + 123.55E + 111.83E + 138.23E + 131.56E + 121.96E + 125.31E + 123.15E + 131.63E + 133.09E + 11
Sunda, E.950f71b145j 7.10E + 112.64E + 123.20E + 122.01E + 133.82E + 111.15E + 123.70E + 121.82E + 138.26E + 121.57E + 11
Tonga1460a148a100a8.51.56E + 1106.22E + 111.24E + 132.36E + 113.59E + 121.08E + 135.70E + 132.74E + 135.20E + 11
Trobriand590f20c50d      1.85E + 114.77E + 112.98E + 121.59E + 123.01E + 10
Vanuatu1189a111a40a7.51.29E + 123.33E + 124.20E + 125.26E + 131.00E + 122.03E + 124.95E + 123.29E + 131.81E + 133.43E + 11
Yap-Palau550c3g32d      2.49E + 105.70E + 104.05E + 112.30E + 114.37E + 09
Total (26 zones)       2.72E + 133.95E + 139.27E + 135.79E + 141.10E + 13     
Global total44454      3.62E + 135.26E + 131.24E + 147.72E + 141.47E + 133.81E + 131.00E + 146.14E + 143.23E + 146.14E + 12

[39] Convergence rates for the major plates are based on the NUVEL-1A global motion model [DeMets et al., 1990, 1994], supplemented by GPS data [Bevis et al., 1995; Tregoning et al., 1998] and the Philippine plate motion of Seno et al. [1993]; some Philippine plate GPS data are too recent to be included [Miyazaki and Heki, 2001; Kato and Kotake, 2002]. Only the component of convergence perpendicular to each trench is considered. When applicable, back-arc spreading rates are added to the major-plate motions, using back-arc rates summarized by Jarrard [1986]. A few convergence rates for short trenches in Southeast Asia (Manila, Cotabato, Negros, Sulu, and Trobriand) are poorly known but slow, and crustal age at San Cristobal is ill-determined; subsequent calculations of global fluxes are insensitive to these local uncertainties. For most of the longest subduction zones, McCaffrey [1997] has computed perpendicular-convergence rates and average subducting-plate ages pointwise along each trench, obtaining a more accurate average for each than is obtainable by subjectively selecting a representative crustal age, trench azimuth, and convergence rate for an entire subduction zone. In general, I use McCaffrey's [1997] averages when available (Table 2), but I have not undertaken a similar pointwise approach for the other subduction zones.

[40] Total consumption of oceanic crust on the subduction zones of Table 2 is 2.42 km2/yr, much less than the 3 km2/yr (3.3 km2/yr including back-arc spreading) of crustal generation calculated from early global motion solutions [Minster and Jordan, 1978; Parsons, 1981] and usually used for estimation of subduction fluxes [Ito et al., 1983; Peacock, 1990; Bebout, 1995; Staudigel et al., 1996]. Part of this difference is due to substantial reductions (10–25% in the Pacific) of many seafloor spreading rates by NUVEL-1, 2% of which is attributable to time-scale revisions [DeMets et al., 1990], with a further 4.4% time-scale-induced reduction by NUVEL-1A [DeMets et al., 1994]. I am not aware of a recalculation of global crustal generation based on NUVEL-1 or NUVEL-1A. In addition, total consumption of oceanic crust is less than total generation because of continent-continent convergence, such as the ∼0.2 km2/yr convergence between India and Asia.

3.2. Sediment Subduction

[41] Some previous analyses of global subduction fluxes have excluded sediment subduction [Staudigel et al., 1996; Alt and Teagle, 1999], and some have included it by assuming a “typical” package of incoming sediments [Ito et al., 1983; Peacock, 1990; Bebout, 1995]. The partitioning of incoming sediments into accreted and subducted components has presented a challenge to quantitative analyses. Hilde [1983] emphasized the role of flexural grabens as sediment traps fostering sediment subduction, and he documented the circum-Pacific occurrences of these grabens. More recent multichannel seismic profiling across accretionary margins has permitted identification of the décollement separating accreting from underthrusting sediments. Three analyses dealt with the extreme regional heterogeneity of sediment subduction by undertaking global syntheses.

[42] Von Huene and Scholl [1991] compiled trench length, pre-NUVEL-1 orthogonal convergence rate, and trench sediment thickness for virtually all modern subduction zones. They used these data, along with reference porosity/depth curves and identification of accretionary versus nonaccretionary margins, to calculate total fluxes of both subducting sediment (1.5 km3/yr, or 4.0 × 1015 g/yr) and pore water (0.9 km3/yr or 9 × 1014 g/yr). Incoming sediment was partitioned between subduction and accretion by assuming 100% subduction for subduction zones lacking prisms, 80% for those with small to medium-sized prisms, and 70% for those with large prisms. von Huene and Scholl's [1991] analysis is the only one that has evaluated trench-fill sediments and subduction erosion. They estimated that the global volume of subduction erosion is comparable to that of deeply subducted incoming sediment, and that the combined total is 1.3–1.8 km3/yr.

[43] Rea and Ruff [1996] used a quite different set of assumptions to assess global sediment subduction. Rather than consider trench-fill, their sediment thicknesses were based on DSDP and ODP reference sites seaward of trenches. This approach allowed them to calculate the amount of incoming sediment for three lithologies (terrigenous, calcite, and opal) and to accurately separate mass fluxes of minerals and pore water. Uneven distribution of reference sites necessitated grouping of many adjacent subduction zones into “trench systems”. Accretion was assumed to be zero for the incoming sedimentary blanket sampled at the reference sites and 100% for the unconsidered trench fill. Rea and Ruff [1996] estimated that the global sedimentary flux of solids is 1.43 × 1015 g/yr and of pore fluids is 9.1 × 1014 g/yr. This fluid flux is virtually identical to that of von Huene and Scholl [1991], whereas the solid flux is only 35% of theirs. The analysis of Rea and Ruff [1996] did not consider fluxes of structural water, K2O, or CO2, but their calcite flux of 2.2 × 1014 g/yr suggests a sediment-hosted CO2 flux of ∼1.0 × 1014 g/yr.

[44] The most comprehensive analysis of sediment subduction is that of Plank and Langmuir [1998], who calculated subduction fluxes for major and trace elements at 26 subduction zones. Like Rea and Ruff [1996], they computed sedimentary components based on analysis of DSDP and ODP reference sites on the incoming plate, but their more detailed lithologic analysis was calibrated with geochemical measurements, permitting calculation of bulk compositions and therefore elemental fluxes. By comparing sediment-hosted elemental fluxes for individual subduction zones to the composition of arc volcanics, they were able to evaluate the effects of sediment subduction on arc volcanism.

[45] The analysis of Plank and Langmuir [1998] quantitatively assessed the sediment-hosted subduction fluxes of four components focused on in this study: structural water (H2O+), pore water, CO2, and K2O. Chloride, the fifth component, is readily calculated from pore water. Their tabulations confirmed that H2O+ and K2O are ubiquitous in subducting sediments: H2O+ occurs mainly in clays but also in opal, and K2O is present in the nonbiogenic, terrigenous component. In contrast, CO2 is often virtually absent from subducting sediments; it is, of course, highest in calcareous units, but it is also moderately abundant in clays and volcaniclastics.

[46] Plank and Langmuir [1998] used more accurate, trench-specific measurements of partitioning between sediment subduction and accretion than any other study. Many of their rates and trench lengths were from von Huene and Scholl [1991] and are now superceded. Applying the revised rates (mostly NUVEL-1A) and trench lengths of this study (Table 2) to Plank and Langmuir's [1998] concentrations of structural water, pore water, CO2, and K2O changes fluxes at most individual subduction zones substantially. Their global fluxes of these four components are quite robust, however, and are revised by only −1.5% to +3%.

[47] A limitation of the Plank and Langmuir [1998] study is that its reliance on reference drill sites precludes analysis of 14 of the subduction zones of this study. The most volumetrically significant omissions are probably the Aegean and the three segments of Chilean subduction zones. Their analysis includes two thirds of the global trench length represented in Table 2, but a more useful correction factor would be percentage of globally subducted sediment included in their analysis. Based on the virtually complete suite of global subduction zones analyzed by von Huene and Scholl [1991], the subset analyzed by Plank and Langmuir [1998] accounts for 75% of total globally subducted sediments. Consequently, I assume that the global sediment-hosted fluxes of structural water, pore water, CO2, and K2O are one-third higher than those calculated by Plank and Langmuir [1998] and slightly refined above. It should be noted, however, that the heterogeneity of sediment CO2 contents makes that one-third adjustment less accurate for CO2 than for the other three components.

3.3. Oceanic Crustal Layers

[48] Oceanic crustal layering was first determined based on seismic refraction. The Raitt-Hill layering [Hill, 1957; Raitt, 1963] consists of the following: Layer 2, the “volcanic layer”, which is 1.71 ± 0.75 km thick; underlain by Layer 3, the “oceanic layer”, which is 4.86 ± 1.42 km thick; separated by Moho from Layer 4, the upper mantle. Many subsequent seismic experiments have refined the seismic structure of Layer 2, its sublayers 2a, 2b, and 2c, and Layer 3. The synthesis of White et al. [1992] concluded that normal Layer 2 is 2.1 ± 0.6 km thick and Layer 3 is 5.0 ± 0.9 km thick, for a total crustal thickness of 7.1 ± 0.8 km.

[49] Attempts to correlate this layering with petrological variations have met with limited success. Moho is a distinctive boundary, but seismic Moho may not correspond exactly with petrological Moho. The deepest drilling into in situ oceanic crust is at Hole 504B. The top layer here is a 572 m extrusive volcanic section, underlain by a 208-m thick transition zone (572–780 m sub-basement), underlain by sheeted dikes that continue to the 1836 m sub-basement bottom of the hole. The top ∼100 m is low-velocity Layer 2A, but the main petrological change within the extrusives is a change from more open to more restricted low-temperature alteration near the middle of the extrusives [Alt et al., 1996b]. The layer 2B/2C boundary, generally considered to mark the change from extrusive basalts to dikes, occurs within the transition zone. The Layer 2C/3 boundary has been commonly considered to be the change from dikes to gabbros [Fox and Stroup, 1981], but at Hole 504B it is probably a metamorphic front within the sheeted dikes [Salisbury et al., 1996].

[50] I follow the example of Hole 504B in assuming that the upper extrusive layer is 600 m thick, divided into two 300-m thick layers that may have significantly different porosity structures and associated alteration states (Table 1). I include the transition zone with the sheeted dikes, because its alteration history is generally similar to that of the upper dikes [Alt et al., 1996b], but its velocity and porosity structure are more like the lower extrusives than the sheeted dikes. I follow many previous investigators in assuming that the base of the dikes is at 2 km sub-basement, while recognizing that this assumption is largely based on the outdated interpretation of Layer 2C/3 at 2.1 km as this petrological boundary. The seismic Moho at 7 km is taken as the base of oceanic crust and the lower boundary of these analyses. These layers are generally much thinner in ophiolites [Coleman, 1977]. An exception is the Semail ophiolite, with crust that is 7 km thick. The seismic structure of this ophiolite is generally comparable to that of normal oceanic crust, but the extrusive section is twice as thick as at Hole 504B, the sheeted dikes extend to 3.3 km, and some peridotites occur above seismic Moho [Christensen and Smewing, 1981].

[51] Crustal accretion at slow spreading rates involves more episodic volcanism, a deeper melt lens [Purdy et al., 1992] and more abundant faulting than that at intermediate-to-high rates, resulting in disrupted igneous stratigraphy and more complex, heterogeneous structure [e.g., Cannat, 1996; Karson, 1998]. Despite these differences, seismic crustal thickness is only weakly correlated with spreading rate [White et al., 1992; Small, 1998], except for rare occurrences of rates <6 mm/yr [Dick et al., 2002].

3.4. Properties of Upper Extrusives

[52] Values for the top 300 m of the extrusive section (Table 1) are based on the core and log analyses of section 2.

3.4.1. Microporosity

[53] Microporosity is a routine shipboard core measurement. The primary control on microporosity of extrusives is volcanic style: pillows have microporosities that are several percent higher than more massive flows. I assume an age-independent microporosity of 7.8% for the upper crust, based on averaging 839 measurements from the normal crustal sites tabulated by Jarrard et al. [2003].

3.4.2. Macroporosity

[54] Macroporosity, consisting mainly of cracks and interpillow voids, decreases systematically with increasing age, as described by the linear regression of Figure 2c. The secondary minerals responsible for this macroporosity reduction are assumed to be an 80:20 mix of saponite and celadonite (section 2.8), the same as at an intergranular scale. Calcite is undoubtedly present as well at a relatively small percentage, which is already accounted for in the analysis of Alt and Teagle [1999] and section 3.4.5 below.

3.4.3. Total Porosity

[55] Total porosity is simply the sum of microporosity and macroporosity.

3.4.4. Matrix Density

[56] Matrix density is assumed to decrease systematically with increasing age, as described in the linear regression of Figure 2b, due to gradual replacement of primary minerals by secondary minerals.

3.4.5. Carbon Dioxide

[57] Carbon dioxide is assumed to increase systematically with increasing age, as described in the linear regression of Figure 2d.

3.4.6. Potassium and Bound Water

[58] Potassium and intergranular bound water are calculated from age-dependent matrix density (Figure 2b), attributing reduction in matrix density to an 80:20 mixture of saponite and celadonite (Figure 5). Bound water is also present within alteration minerals filling fractures and interpillow voids (section 3.4.2); these alteration minerals are also assumed to be a saponite/celadonite mix.

3.5. Properties of Lower Extrusives

[59] Significant sections of the lower extrusives (300–600 m sub-basement) have been penetrated at only four sites: Holes 395A, 418A, 504B, and 801C. Consequently, generalizations concerning the properties of this zone are based primarily on comparisons of these four sections to their overlying extrusive zones. Tectonic windows provide another perspective on this zone [e.g., Karson, 2002], if they are representative [Francheteau et al., 1975].

[60] The upper and lower extrusives are expected to be broadly similar in geophysical characteristics, volcanically and hydrologically. Volcanic style may exhibit recurrent patterns of flows, pillows, and hyaloclastics [Robinson et al., 1979; Schmincke and Bednarz, 1990; Pezard, 1990; Pezard et al., 2000], but resulting stratigraphic architecture has a wavelength of tens of meters, not several hundreds of meters. Large-scale permeability measurements of upper oceanic crust are too sparse for quantitative comparison of upper versus lower extrusives, but the first-order generalization of these studies is that the entire extrusive interval acts as a hydrologic unit during off-axis hydrothermal circulation [Fisher, 1998; Fisher and Becker, 2000]. Seismic studies (e.g., ESP, OBS, OBH) indicate strong velocity gradients within the top few hundred meters of oceanic crust [Vera and Mutter, 1988; White et al., 1992], whereas no systematic pattern of downhole velocity increase is detected within the upper 300 m of basement for DSDP and ODP sites [Jarrard et al., 2003].

3.5.1. Core-Based Alteration

[61] Core descriptions of the four deep-crustal sites provide ground-truth comparisons of variations in alteration with depth. In all four extrusive sections, smectite is the most abundant alteration mineral. At Hole 395A, depth-dependent alteration was noted principally as downhole increases in non-carbonate veins [Shipboard Scientific Party, 1978] and temperature of mineral formation [Lawrence et al., 1978], though nearly all diagenesis is low-temperature. At Hole 418A, alteration is consistently higher in the top 200 m than in the underlying 250 m [e.g., Broglia and Ellis, 1990]. At Hole 504B, the upper and lower extrusives are generally similar in alteration mineralogy, but the lower extrusives are less altered [Alt et al., 1996b] and have more chlorite and mixed chlorite-smectite [Alt et al., 1985, 1986]. The tholeiitic section at Hole 801C penetrates only ∼100 m of the lower extrusives, and depth-dependent changes in alteration are difficult to isolate from flow/pillow-induced heterogeneity (Figure 4).

[62] Chemical analyses generally underestimate alteration because of sampling bias and underrecovery of both altered basalts and macroporosity-filling alteration minerals. For example, despite the observed difference in alteration between upper and lower extrusives at Hole 504B [Alt et al., 1996b], H2O+ analyses are similar: 0.86% (std. dev. 0.32%, N = 51) and 0.83% (std. dev. 0.66%, N = 77), respectively. The short interval of lower extrusives at Hole 801C continues an overall pattern of downhole decrease in alteration, particularly decrease in CO2 and possibly also H2O+ but not K2O [Shipboard Scientific Party, 2000a]. Staudigel et al. [1996] quantitatively distinguished recovered and unrecovered materials for Site 417/418, by carefully weighting their samples to achieve representative sampling. Like some core/log comparisons, their analysis demonstrated that total abundance of alteration minerals is much higher than most XRF measurements detect. In the lower extrusives, average K2O is 14%, structural H2O is 64%, and CO2 is 19% of that in the upper extrusives [Staudigel et al., 1996].

3.5.2. Log-Based Alteration and Total Porosity

[63] Potassium data from cores indicate that most K-rich alteration minerals are in the uppermost crust, and they suggest a near-exponential decrease to ∼600 m sub-basement [Booij and Staudigel, 1997]. Potassium logs provide unbiased, integrated comparisons of total potassium enrichment for upper versus lower extrusives at Holes 395A, 418A, and 504B. The four potassium logs for Hole 395A indicate that the lower extrusives have 110%, 101%, 120%, and 128% of the K2O content of the overlying extrusives. In contrast, the potassium log for the lower extrusives at Hole 418A averages only 57% of the K2O in the upper extrusives. At Hole 504B, the corresponding proportion is only 35% based on the potassium log and 46% based on 458 XRD measurements (Figure 3).

[64] All three sites have both multichannel sonic and reprocessed Schlumberger sonic logs that yield similar large-scale velocities; in conjunction with core-based velocities, these provide macroporosity estimates. In general, the deeper extrusives have macroporosities that are about half those of the shallower extrusives: 0–3.7% versus 6.0–8.2% at Hole 395A, 1.8–2.1% versus 3.4–4.2% at Hole 418A, and 2.5–5.0% versus 8.0–8.1% at Hole 504B. Similarly, core-based microporosities for Holes 395A, 418A, 504B, and 801C are significantly lower for the lower extrusives than upper ones, averaging 5.1% (std. dev. 3.8%, N = 154) and 7.8% (std. dev. 5.9%, N = 839), respectively (medians 4.0% and 5.8%). Matrix densities are different as well: 2.97 g/cc for lower extrusives versus 2.93 g/cc for upper ones, corresponding to 8% and 16% alteration, respectively, based on the mineral alteration model of section 2.8.

[65] In summary, the bulk of the available evidence indicates that the lower half of the extrusive layer has about half as much macroporosity as the upper half. This difference is apparently not alteration induced; indeed, matrix densities and potassium logs suggest about half as much hydration and potassium enrichment for the lower extrusives as for the upper ones. The weight of the volcanic pile may cause a few percent more macroporosity reduction of the lower extrusives compared to upper ones, thereby reducing permeability and associated fluid flow and generating a seismically detectable vertical velocity gradient. Seawater that does eventually penetrate the lower extrusives generates reducing conditions, lacking the celadonite of the upper extrusives but retaining the strong smectite dominance.

[66] The flux estimates of this paper assume that the lower extrusives exhibit age-dependent macroporosity and alteration that are half those of the overlying extrusives. Age-dependent changes in matrix density, structural water, carbon dioxide, and potassium are calculated from this assumption of proportional alteration (Table 1).

3.6. Properties of Dikes

[67] Hole 504B could be the type section for Layer 2 of oceanic crust. Seven DSDP and ODP legs at this site have cored and logged through a 572-m thick extrusive layer, a 208-m transition zone, and 1056 m of sheeted dikes, thought to represent virtually the entire dike layer. No other DSDP or ODP site has sampled the dike layer (Layer 2C and upper Layer 3), though isolated dikes have been encountered within the extrusive layer. The only other stratigraphic sections of this portion of oceanic crust are from ophiolites [e.g., Coleman, 1977; Christensen and Smewing, 1981], whose geophysical properties may have been affected by emplacement and post-emplacement processes. Consequently, this study bases its characterization of the dike layer on Hole 504B, combining the transition zone and sheeted dikes into a nominally uniform layer.

[68] Alt et al. [1996b] provided a synthesis of alteration mineralogy and the multistage alteration history of the Hole 504B dikes. Olivine alters to chlorite, clay, and talc, plagioclase to albite, zeolites, chlorite, or clay, and augite rarely alters to actinolite [Alt et al., 1996b]. Alteration of the upper dikes and transition zone began with greenschist-facies metamorphism at 350–380°C during axial hydrothermal alteration, followed by mixing of upwelling hydrothermal fluids with cooler seawater circulating through the extrusives. Alteration of the lower dikes began with exposure to very high-temperature (>400°C) hydrothermal vent fluids, followed by 300–400°C greenschist metamorphism, and concluding with rare off-axis (<250°C) veins [Alt et al., 1996b].

[69] Index measurements on 91 samples from the transition zone and sheeted dikes of Hole 504B establish a mean matrix density of 2.98 g/cc (standard deviation 0.04 g/cc) and mean microporosity of 2.25% (standard deviation 2.10). Macroporosity calculation based on comparison of lab and log velocities must use lab measurements at elevated pressure rather than at atmospheric pressure, because microcrack opening becomes significant at depths >300 m into the dikes. With this approach, Salisbury et al. [1996] estimated dike macroporosities as 0–1.4%. Average macroporosity of the dikes is 0.8%, based on lithostatic-pressure core measurements. Most of this macroporosity is concentrated in the transition zone, where macroporosity is 5.6%, in contrast to near-zero (mean −0.1%) in the sheeted dikes.

[70] Concentrations of structural water are relatively uniform within the dikes of Hole 504B, except for significantly higher values within the transition zone (mean 2.37%, std. dev. 0.78%, N = 140) than in the sheeted dikes (mean 1.67%, std. dev. 0.67%, N = 274). I assume that the overall average of 1.76% (std. dev. 0.72%, N = 314) is representative of Layer 2C globally. Unlike the situation for the extrusives, this value is probably not significantly biased by loss of structural water during sample preparation, because pure smectites are absent and expandable clays are not abundant [Alt et al., 1985, 1996b]. Although core recovery in the dikes is generally low, the core-based determination of structural water may not be significantly biased by preferential coring loss of crack-fill alteration minerals, because H2O+ measurements suggest relatively uniform hydration. Broglia and Ellis [1990] analyzed a neutron log of upper dikes (875–1250 m sub-basement) of Hole 504B, accounting for borehole effects, porosity, and structural water contributions. They concluded that volume percentage of total water is 7–8% within the upper dikes; of this, porosity is <2% and is probably entirely microporosity, whereas hydrous mineral content of about 10–30% contributes 6% H2O.

[71] Considering both whole-rock CO2 and vein carbonate, average CO2 at Hole 504B is 0.58 wt.% for the transition zone and 0.07% for the sheeted dikes [Alt and Teagle, 1999]. The thickness-weighted average is 0.14%.

[72] K2O content of the dikes of Hole 504B has been measured via XRF analyses of cores and downhole logging (Figure 3). K2O leaching [Alt et al., 1996b] has so depleted this interval, however, that it is near or at the resolution of both methods. Based on 343 XRF measurements, average K2O is 0.014% (std. dev. 0.013), with values that decrease gradually downhole (Figure 3) from a level of 0.023% (std. dev. 0.011) in the transition zone to 0.011% (std. dev. 0.012) in the underlying sheeted dikes. Spectral gamma ray logging indicates significantly higher K2O: 0.034% for ACT logging of 1075–1823 mbsf on Leg 148, and 0.057% for logging of 1349–1871 mbsf on Leg 140. The difference between core and log K2O could be caused by either non-representative core recovery (as observed within the extrusives) or measurement bias. I attribute the difference to biased log measurements, resulting from K2O levels that are much lower than those anticipated during tool design. For example, minimum K2O detectability and minimum K2O readings are both sensitive to trace amounts of K2O in the detector crystal. Different detector crystals may account for the 65% difference in average K2O for logs obtained on different legs. I assume that K2O is sufficiently uniform within the dikes (Figure 3) for the mean XRF-based K2O of 0.014% to provide a representative estimate of the K2O content of Layer 3. This low value results from leaching of an already K2O-depleted magmatic composition [Alt et al., 1996a, 1996b]. More typical MORB compositions may or may not have higher final K2O contents; K2O availability for leaching may determine final K2O content (J. Alt, personal communication, 2002).

[73] With only one dike section available, it is impossible to examine the possibility of age-dependent dike alteration directly. Because the alteration mineral assemblage within the Hole 504B dikes is mainly high-temperature [Alt et al., 1996b], it is largely attributable to near-ridge alteration. Consequently, I assume that dike alteration is age-independent, at least for the 5–145 Ma average ages of subducting crust in modern subduction zones (Table 2). The heat flow pattern at Hole 504B suggests present-day hydrothermal circulation within the dikes [Wang and Davis, 2002], and late-stage dike alteration may be occurring [Alt et al., 1996b], so the assumption of age-independent alteration could be only a first-order approximation. I apply the Hole 504B values for dike microporosity, macroporosity, structural water, K2O, CO2, and matrix density to the dike layers at all subduction zones.

3.7. Properties of Gabbros

[74] The processes that bring the gabbro layer within sampling reach also assure that the samples will be nonrepresentative of normal in situ conditions. DSDP and ODP have encountered short gabbro intervals at various sites, often where only basalts were anticipated. Significant penetrations of Layer 3 gabbros were obtained at two localities: Hole 735B on Atlantis Bank of Southwest Indian Ocean Ridge [Shipboard Scientific Party, 1989, 1999] and Site 894 at Hess Deep [Shipboard Scientific Party, 1993]. At both localities, faulting has unroofed the gabbroic section and exposed it at the seafloor, resulting in stratigraphic sections that are enhanced in fracturing, microcracks, and low-temperature alteration, compared to what is likely for normal in situ Layer 3 [Shipboard Scientific Party, 1989, 1993, 1999; Bach et al., 2001]. Nevertheless, I assume that these two sites provide the most representative sample of Layer 3 currently available.

[75] Hole 735B was cored to 501 mbsf on Leg 118 and deepened to 1508 mbsf on Leg 176. Core recovery was exceptionally good, but loss of the bottomhole assembly during Leg 176 prevented downhole logging of much of the newly drilled interval. The entire drilled interval consists of Layer 3 gabbros. Index measurements on 346 samples establish a mean matrix density of 2.99 g/cc (standard deviation 0.12 g/cc) and mean microporosity of 0.74% (standard deviation 2.76). Ildefonse and Pezard [2001] measured 63 additional samples, which have a mean matrix density of 2.98 g/cc and microporosity of 1.11%. The microporosity distribution is, however, non-normal, skewed positively by a few high-porosity samples; median microporosity of the 346 samples is only 0.2%. This microporosity is partly microcracks, based on resistivity measurements [Ildefonse and Pezard, 2001] and on velocity measurements that average 5.6% higher at 200 MPa than at atmospheric pressure [Stephen, 2001]. Microcracks may not be present in unexhumed gabbros of normal oceanic crust.

[76] Amphibolite-facies alteration of upper Hole 735B is extensive, averaging 20–30% and with veins comprising 2.4% of total volume [Dick et al., 1991; Robinson et al., 1991]. By comparison, alteration of the Semail ophiolite is characterized by rare conversion of clinopyroxene to actinolite in the layered plutonics [Lippard et al., 1986] and by actinolite and Fe-Ti oxides plus cavity fillings in more isotropic plutonics [Ernewein et al., 1988].

[77] Measurements of structural water at Hole 735B are most abundant for the interval drilled during Leg 118: 0.79% average of 94 samples [Shipboard Scientific Party, 1989]. I assume that this value is representative of gabbros in general, but high-temperature alteration of gabbros may generally be limited to the upper few hundred meters by available fracture permeability [Alt, 1995]. At Hole 735B, high-temperature alteration extends to about 600 mbsf, and unroofing-induced low-temperature alteration is locally common below 500 mbsf [Bach et al., 2001]. Consequently, the 1.97% average structural water concentration for 20 altered samples from below 500 mbsf is less representative than the 0.79% average for 18 fresh samples [Bach et al., 2001]. The gabbros cored at Hess Deep Site 894 average 1.26% for 22 samples (1.39% including one extremely hydrated sample). Carlson [2001] cited an average bound water content of only 0.04% for the gabbros of Holes 735B, 894G, and 923A.

[78] CO2 content of gabbros is assumed to be 0.02 wt.%, the average of whole-rock analyses from upper Hole 735B, Hess Deep, and MARK sites [Alt and Teagle, 1999]. Fresh gabbros from the deeper portion of Hole 735B, however, have a much higher median CO2 of 0.09% [Bach et al., 2001].

[79] K2O content of the gabbros of Hole 735B has been measured via XRF core analyses and downhole logging. Based on 280 XRF measurements from Legs 118 and 176 [Shipboard Scientific Party, 1989, 1999], average K2O is 0.065% (0.061% excluding a single extreme value of 3.97%). Leg 176 samples affected by low-temperature alteration average twice as much K2O as fresh samples [Bach et al., 2001]. Three spectral gamma ray logs of the upper portion of this interval give reasonably consistent average K2O contents (0.124%, 0.138%, and 0.146%) that are twice as high as XRF results. The good log replicability and mean K2O readings that are apparently well above measurement threshold argue against log measurement bias as an explanation. The logs may provide a more representative sample than cores of the K2O contents of gabbros at Hole 735B, yet be less appropriate for evaluating the K2O of gabbros in general. The logs are more sensitive to occasional K2O spikes caused by low-temperature alteration minerals, present at Hole 735B because of its unroofing history but not expected to be present in deep gabbros.

[80] I assume that Layer 3 gabbros have an average K2O of 0.061%, based on the Hole 735B XRF results; this value is close to the weighted average of 0.05% determined by Dick et al. [2002]. This value is significantly less than the 0.11% average for the upper third of this hole based on strip samples [Hart et al., 1999]. It is consistent with the XRF-based average of 0.08% for the much shorter interval of gabbros sampled at Hess Deep Site 894. It should be noted, however, that uncertainty in gabbro K2O has a huge impact on subsequent calculations of total K2O flux into subduction zones: using the log-based value, instead, would increase computed total K2O flux by 71%.

4. Subduction Fluxes

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

4.1. Water Flux

[81] The annual global flux of subducted water is 1.83 × 1015 g/yr. This flux includes four major sources (Figure 6): sediment pore water (42%), sediment structural water (7%), igneous crustal pore water (18%), and igneous crustal structural water (33%). The global contributions from subducting sediments and crust are nearly equal. Regionally, in contrast, combined structural and pore water of incoming igneous crust is nearly uniform, whereas the quantity of sediment pore water subduction per meter of trench length varies by more than an order of magnitude. Variations in total-water fluxes among individual subduction zones (Figure 6) are most sensitive to convergence rate, trench length, and thickness of the subducting sedimentary section.


Figure 6. Water fluxes at global subduction zones, from Table 2. H2O+ is bound water in minerals; for crustal H2O+, loosely bound water is distinguished from water in stabler, more temperature-resistant minerals.

Download figure to PowerPoint

[82] The computed water contents for crust of various ages are remarkably uniform: their total range is only 5.5–5.7%, a variation of only 3% among incoming crust at modern subduction zones. This similarity arises from two factors. First, all age-dependent changes in macroporosity and alteration within dikes and gabbros are assumed to occur near-ridge, so that the structural and pore water contents of these portions of the crust are the same throughout the 5–145 Ma average ages at modern subduction zones. Age-dependent changes in macroporosity and alteration are confined to the extrusives, which account for only 16–31% of total crustal structural water and 35–51% of total crustal pore water. Second and more important, age-dependent decreases in upper crustal macroporosity are nearly balanced by increases in structural water. Most of this structural water is within the mineral matrix, mainly replacement of feldspar and olivine by saponite; only 17% of macroporosity fill is structural water. Although the average structural water content of the extrusive section varies by a factor of 2.3 (2.2–5.1%) among subduction zones, and average pore water content varies by a factor of 1.9 (2.6%–5.0%), leading to an 18% regional variation in total bound water and 24% variation in total pore water, total water varies by only 3%.

[83] The mass balance for water addition and removal within oceanic crust appears to permit off-axis crustal alteration to be confined to the oceanic crust, without significant interchange with seawater, such as might be expected when a relatively impermeable sediment blanket intervenes. Similarly, much of late-stage intergranular alteration is accomplished by local redistribution of elements without substantial bulk-rock geochemical change [Alt, 1999; Zhou et al., 2001a]. However, interchange with seawater to at least ∼65 Ma is indicated by both heat flow data and by the other elemental mass fluxes of this study. The structural water enrichment at the expense of pore water implies either chlorine outflow or increase in pore water chlorinity, and ongoing additions of both potassium and CO2 have been documented earlier.

4.1.1. Comparison to Prior Flux Estimates

[84] My calculated global fluxes of 1.24 × 1014 g/yr for structural water and 7.72 × 1014 g/yr for pore water in subducting sediments are based on minor revisions of the comprehensive evaluation of Plank and Langmuir [1998]. An earlier value of 0.7 × 1014 g/yr for structural water by Peacock [1990] was based on a single hypothetical subducting sediment column, in contrast to the evaluations for individual subduction zones by Plank and Langmuir [1998]. Most prior estimates for sediment pore water flux are slightly higher than my 7.7 × 1014 g/yr: 1 × 1015 g/yr [COSOD-II, 1987], 1 × 1015 g/yr [von Huene and Scholl, 1991], 1.8 × 1015 g/yr [Moore and Vrolijk, 1992], and 9 × 1014 g/yr [Rea and Ruff, 1996]. The good agreement of Rea and Ruff's [1996] value with Plank and Langmuir's [1998] result is partly attributable to methodological similarities in the two analyses, including use of many of the same reference sites.

[85] Most previous analyses of water budgets for oceanic crust do not quantify pore water; they focus instead on structural water. Estimates for the latter are higher than the 6.14 × 1014 g/yr of this study: 8.8 ± 2.9 × 1014 [Ito et al., 1983] and 8 × 1014 [Peacock, 1990]. Ito et al. [1983] used assumed abundances and associated water contents for hydrous alteration minerals in an oceanic crust consisting of alteration-based layers: halmyrolysis, greenstone, amphibolite, and unaltered gabbro. Their structural-H2O values are similar to ours for the extrusives, higher for the dikes, and much lower for the gabbros. Peacock [1990] assumed 2% structural H2O for basalts and 1% for gabbros, based on consideration of whole-rock analyses from three upper-crustal sites and the gabbros of Hole 735B and Oman. Peacock [1996] noted that dredged rocks indicate much higher structural water: 1.7–4.9% (avg. 3.5%) for basalts and 1.6–5.8% (avg. 2.5%) for gabbros, based on the data of Anderson et al. [1976]. Bebout [1995, 1996] suggested that total water flux may be ∼18 × 1014 g/yr, about double the structural-H2O values of Ito et al. [1983] and Peacock [1990], to account for the high H2O+ contents of outcropping metamorphosed subduction-zone rocks such as the Catalina Schist. He noted, however, that this additional water may come from sedimentary or basaltic pore fluids. My structural-H2O flux for the extrusives, 1.58 × 1014 g/yr, is significantly higher than the 1.01 × 1014 of Staudigel et al. [1996], due to a combination of higher structural H2O, thicker extrusive section (600 m versus 500 m), and lower porosity (7–13% versus 18.5%).

[86] Moore and Vrolijk [1992] calculated global water fluxes for both subducting sediments and oceanic crust, using trench lengths and convergence rates from von Huene and Scholl [1991], upper crustal porosity and hydrous minerals for Hole 504B [Becker et al., 1990], and hydrous minerals in sediments of Kastner et al. [1991]. They estimated fluxes (typo-corrected) of 1.8 × 1015 g/yr for sediment pores (versus my 7.7 × 1014 g/yr), 4.3 × 1014 g/yr for H2O+ in sediments (versus my 1.2 × 1014 g/yr), and 3.5 × 1014 g/yr for pore and structural water in the top 1 km of oceanic crust (versus my 3.7 × 1014 g/yr for the same portion).

[87] Nearly all of the water that enters subduction zones comes from seawater or – in the case of bound water in sediments – terrestrial hydrous minerals. Seven studies of fresh MORB glasses indicate about 0.15–0.25% H2O+ [Ito et al., 1983]. Sobolev and Chaussidon [1996] argued that fluid inclusions in olivine phenocrysts provide a more reliable measure of initial MORB H2O+ (0.12 wt.%) than glass does. This value, in conjunction with a crustal consumption rate of 2.4 km2/yr, basaltic layer thickness of 2 km, and weighted average density of 2.76 g/cc, indicates that 1.6 × 1013 g/yr of subducted water, or 0.9% of the total subducted water, is primary water from crustal generation.

4.1.2. Slab Fluid Expulsion Pattern

[88] Subducting fluids control first-order structural and petrologic problems: structural style and evolution of accretionary prisms [Davis et al., 1983] and generation of arc magmas [Gill, 1981] and consequently long-term growth of continents [e.g., Reymer and Schubert, 1984]. Unfortunately, even rough fluid budgets demonstrate that our knowledge of slab fluid expulsion patterns still faces first-order uncertainties: present-day fluid expulsion rates from some subduction zones are an order of magnitude higher than fluid sources [Le Pichon et al., 1991; Kastner et al., 1991], and an order of magnitude more fluid enters subduction zones than is eventually released within arc magmas [Ito et al., 1983]. Yet, the 1.83 × 1014 g/yr loss of water to subduction zones must be nearly balanced by gains elsewhere, because even a sustained 20% imbalance implies a long-term sea level change of ∼1 m/Ma, which is unlikely on the time scale of >100 Ma.

[89] Figure 7 summarizes slab fluid expulsion patterns, based on the incoming global water budgets of Table 2 and dehydration evidence discussed in this section. A useful starting point is to stipulate that there are five types of subducting water, each of which might be released at a different point in the subduction process: sediment pore water (7.7 × 1014 g/yr), sediment structural water (1.2 × 1014 g/yr), igneous-crust pore water (3.2 × 1014 g/yr), igneous-crust loosely bound structural water (1.6 × 1014 g/yr), and igneous-crust firmly bound structural water (4.6 × 1014 g/yr). A likely sixth source, structural water in serpentinites, is ignored here because its magnitude is unknown. Structural water within the extrusive portion of oceanic crust is mostly in smectites, whereas that in the dikes and gabbros is in more temperature-resistant minerals such as actinolite and hornblende, so release of the former may be more closely related to smectite breakdown within sediments than it is to breakdown of lower crustal hydrous minerals. Structural water within the sediments is primarily in clays (particularly smectite) and secondarily in opal, and most smectite interlayer water is released at temperatures below 150°C.


Figure 7. Hypothesized water expulsion pattern in a generalized subduction zone. Bottom: cross-section of a subduction zone, with solid arrows showing escape paths of subducting water, and dashed lines showing mantle flow paths. Dehydration mechanisms are shown above the cross-section, with solid bars above the portion of slab undergoing each dehydration reaction. The plot of total global subducted water semiquantitatively illustrates expulsion magnitudes; starting values are from Table 2, but expulsion rates are only approximate, as discussed in text.

Download figure to PowerPoint

[90] Subduction at margins containing accretionary prisms expels water due to compaction within the accreting and underthrust sediments [Bray and Karig, 1985] and transformation of smectite to illite. Kastner et al. [1991] and Le Pichon et al. [1991] simultaneously demonstrated the discrepancy between present rates of water expulsion at Barbados, Nankai, and Peru prisms and the amount of water that can be accounted for by generation within the prism. Kastner et al. [1991] calculated average water expulsion of ∼7 m3/yr per m of trench (global mass of ∼3 × 1014 g/yr) from internal fluid sources (compaction and dehydration), much less than the 100 m3/yr per m of trench presently being vented at the three margins. The difference was initially thought to be provided mainly by meteoric water [Kastner et al., 1991] and shallow seawater convection [Le Pichon et al., 1991].

[91] Isotopic studies and identification of low-chlorinity and high-methane anomalies demonstrated the significant contribution from deeper diagenetic processes, particularly smectite dehydration and hydrocarbon generation [Kastner et al., 1991, 1993]. In addition, the assumption of steady state compaction may be invalid [Moore and Vrolijk, 1992; Le Pichon et al., 1993]. Recent modeling of flow and solute transport at Nankai accretionary prism incorporated transient flow, caused by a hydrofracture-induced temporary increase in permeability along the décollement [Saffer and Bekins, 1998].

[92] Subducted sediments are initially overpressured and underconsolidated [Davis et al., 1983; Moore and Vrolijk, 1992]; rate of fluid expulsion depends on a combination of loading, convergence rate, and permeability [Saffer and Bekins, 2002]. Normal consolidation is achieved and most smectite dehydrates after perhaps 30–40 km of subduction (∼5 km burial) [Saffer and Bekins, 1998; Moore and Saffer, 2001]; near-complete loss of pore and structural waters is presumed to occur within a few additional kilometers of burial. It appears doubtful, however, that any significant net escape of fluids from oceanic crust occurs at these shallow depths, because the framework strength of basalts prevents compaction. This does not preclude fluid flow within the basalts, possibly including flushing of water whose chlorinity has been lowered by saponite breakdown, driven by the buoyancy force of warmer waters flowing updip from deeper sources. Vein and fabric studies of outcropping forearc sedimentary rocks indicate an evolution of fluid flow style during subduction: initial mud-filled veins, then cracks with calcite fill, then scaly fabric [Fisher, 1996]. Updip migration of fluids may occur mainly within the high-permeability scaly fabric of the upper portion of underthrust sediments, at least for burial depths of <15 km [Fisher, 1996].

[93] Studies of the Catalina Schist showed a slab-parallel melange fabric, suggesting that most fluid loss at depths of 15–45 km is similarly updip, toward the toe of the prism [Bebout, 1991]. At these depths, fluid loss from subducted sediments is nearly completed and slab devolatilization begins, driven by metamorphism rather than compaction. Bebout [1995] documented this process for the metasedimentary and metamafic rocks of the Catalina Schist, including loss of both water and CO2 and associated homogenization of oxygen and hydrogen isotopic signatures. Bebout [1996] extended these observations to other eclogite-facies subduction complexes. For both metasedimentary and metamafic rocks of the Catalina Schist, volatile loss progresses with increasing metamorphic grade, from H2O+ concentrations of 5–6% for lawsonite-albite to 1–2% for amphibolite-facies metamorphism [Bebout, 1995]. Bebout [1995] noted that H2O+ contents of up to 10% for the low-grade metabasalts are higher and more homogeneous than those of DSDP/ODP basalts and ophiolites. He suggested that low-grade metamorphism transforms both original hydrous minerals and pore water into new hydrous minerals.

[94] Although the Catalina Schist is unlikely to be representative of all subducted sediments and basalts, its pattern of progressive water loss provides a first-order indication of devolatilization at moderate depths. By ∼15 km depth, a water content of 5% by weight for sediments, for a global sediment subduction of 1.7 × 1015 g/yr, implies 8.4 × 1013 g/yr of water subduction, only 9% of the pore and structural water within initially subducted sediments. By ∼40 km depth, reduction of this water content to ∼1.3% expels all but 2% of the initial water. Assuming that the Catalina Schist metabasalts are representative of the extrusive portion of oceanic crust, 5% water content at ∼15 km depth implies 2.1 × 1014 g/yr of water, or 72% of the originally subducted upper crustal waters, and 1.3% at ∼40 km depth implies a residual water content there of 5.6 × 1013 g/yr, or 19% of original subduction volumes (Figure 7). By ∼15 km depth, 50% of the water that entered the subduction zone has been expelled, mainly updip toward the prism toe. By ∼40 km, 60% has been expelled. Inclusion of water expulsion by metamorphic reactions within the dikes and gabbros presumably would raise this total substantially. For example, similar amphibolite-grade metamorphism of the dikes, reducing their water content from 2.8 to 1.3%, would expel 1.5 × 1014 g/yr, or an additional 8% of originally subducted water.

[95] Magnetotelluric surveys have detected high-conductivity zones attributed to free water in or near the top of the subducting slab at several subduction zones: at 25–50 km depth in Juan de Fuca [Wannamaker et al., 1989], at roughly 20–60 km beneath the Okinawa Trough forearc and somewhere within the interval 60–130 km depth beneath the arc [Shimakawa and Honkura, 1991], extending to >60 km depth in Izu-Bonin [Toh, 1993], and from 13 km to 30 km depth in Mexico [Arzate et al., 1995]. The excellent resolution of the Juan de Fuca resistivity model detected an order-of-magnitude conductivity decrease from the trench to 60 km inland, due to pore water loss, followed by a sudden increase in conductivity at 25 km depth, interpreted as dehydration of greenschist-facies minerals [Wannamaker et al., 1989].

[96] Updip reflux persists at least to 15 km and possibly to 45 km, based on the Catalina Schist melange fabric [Bebout, 1991, 1995]. The Juan de Fuca conductivity increase at 25 km [Wannamaker et al., 1989], however, does not suggest reflux. Whether subducting crust below 45 km depth retains sufficient fracture permeability for volatile escape is unknown. Dehydration may generate its own permeability, by inducing hydrofracture [Fyfe, 1997]. A magnetotelluric survey of the Vancouver Island portion of Cascadia subduction zone detected porosities of 0.5–4.0% (assuming seawater salinity) well above the slab, attributed to fluid rise from the slab to an impermeable barrier [Hyndman, 1988].

[97] When the slab reaches 80–120 km depth, large-scale release of volatiles into the overlying mantle wedge generates the partial melting that results in arc volcanism [Gill, 1981]. Except perhaps for very young slabs, the crust undergoes little partial melting [Defant and Drummond, 1990; Peacock, 1990; Peacock et al., 1994].

[98] Based on a variety of H2O+ studies of arc magmas, Ito et al. [1983] calculated that the water flux at arcs is ∼1 × 1014 g/yr; major uncertainties in magma volumes result in a possible range of 0.95–2.0 × 1014 g/yr. Peacock [1990] provided a similar estimate, 1.4 × 1014 g/yr, without mentioning data sources. A more appropriate measure of the volatile content of magmas may be fluid inclusions in olivine phenocrysts, which average 2.5% [Sobolev and Chaussidon, 1996]; in conjunction with an arc magmatism rate of 2.9–8.6 km3/yr [Crisp, 1984], resulting water flux is 2–6 times that of Ito et al. [1983] or Peacock [1990]: 2–6 × 1014 g/yr. Of the ∼6 × 1014 g/yr of subducting water that appears to reach ∼45 km depth, at least a third and perhaps all is later released to the mantle wedge and reaches the arc. The remainder may be released into other parts of the mantle wedge, enter arc magmas but degas during volcanism, or be retained into the deep mantle. I emphasize, however, that the cumulative errors in this mass balance are huge and impossible to quantify. For example, Reymer and Schubert [1984] calculated an arc magmatism rate of 1.1 km3/yr (1.3 km3/yr with a more accurate total trench length) from seismic profiles across arcs that is only 13–38% of Crisp's [1984] rate from active arc volcanism; I use the latter. The unknown amount of water loss to the deeper mantle makes estimation of the mantle water budget problematic [Bell and Rossman, 1992; Thompson, 1992; Williams and Henley, 2001].

[99] Many authors have interpreted phase diagrams based on experimental petrology to infer the cause of the fluid release responsible for arc magmatism [e.g., Peacock, 1990, 1991, 1993, 1996; Pawley and Holloway, 1993; Pawley, 1994; Poli and Schmidt, 1995; Liu et al., 1996; Iwamori, 1998, 2001; Ono, 1998; Ernst, 1999]. A single volatile-expulsion event is not expected. A suite of hydrous minerals (lawsonite, chlorite, amphibole, epidote/zoisite, chloritoid, and others) is expected to break down in a series of overlapping depth zones [Schmidt and Poli, 1998; Kerrick and Connolly, 2001]. Systematic lateral variations in water-soluble elements in magmas from forearc to arc suggest a progressive decrease in water input from the slab [Ryan et al., 1996], possibly a change from amphibole breakdown near the trench to phlogopite decomposition in the backarc [Tatsumi and Kogiso, 1997].

[100] Even if a single volatile-release event were likely within each slab, it should not occur at similar depths in subduction zones with widely varying slab ages and subduction rates, based on pressure-temperature phase diagrams. Slab temperature at any depth is very sensitive to both subduction rate and slab age, as confirmed by thermal models. For example, for a constant subduction rate, the temperature of the top of upper crust at 100 km depth varies from 540°C for 145-Ma crust to 1180°C for 5-Ma crust [Peacock, 1990]. Accordingly, the same dehydration reaction that is expected to release water beneath the arc for intermediate geotherms releases that water beneath the forearc for high-temperature geotherms (slow subduction and/or very young crust) and retains that water to the deep mantle for low-temperature geotherms (rapid subduction and/or old crust) [Kerrick and Connolly, 2001]. Early release of volatiles from young, hot slabs may affect compositions of later, deeper arc magmas, based on a possible association between slab age and magma composition in only two subduction zones [Green and Harry, 1999; Harry and Green, 1999]. In contrast, ultrafast subduction, such as occurs today in Tonga, favors retention of some water to the deep mantle, possibly generating chemical anomalies in the mantle [Staudigel and King, 1992].

[101] Volatile release within the subducted slab may occur earlier within the extrusives than in the dikes and gabbros, not only because of more loosely held water in less stable minerals, but also because of larger-scale metamorphic phenomena. Hacker [1996] concluded that conversion of the extrusives to eclogite is complete by ∼250°C, whereas eclogite conversion of the gabbros usually is complete by 550°C and may continue to >800°C.

[102] In contrast to the wide variability of depths for volatile release, both within and among slabs, volatile escape culminating in generation of arc magma is generally confined to ∼100 km depth. Sub-arc slab depths are readily computed from slab intermediate dip (0–100 km) and arc-trench gap, both tabulated by Jarrard [1986] for 26 of the subduction zones of Table 2. All but two sub-arc depths are between 80 and 120 km. In modern subduction zones, subduction rates vary by a factor of 30, and average ages of the slab vary from 5 Ma to 145 Ma, yet no correlation is found between sub-arc depth and subduction rate, slab age, or their product, strongly implying that melting within the mantle wedge is insensitive to temperature of the adjacent oceanic crust. This observation is consistent with thermal models incorporating convection of the mantle wedge, whereas models lacking convection exhibit strong slab-induced cooling of the mantle wedge [Anderson et al., 1980]. Nor does fluid availability affect the depth of melting: sub-arc depth is not correlated with CO2 or any of the four types of water flux, nor with these fluxes per meter of trench length.

[103] Predictions of multiple volatile expulsions and thermally dependent expulsion imply large-scale transport prior to release beneath the arc. Even with available water to lower melting temperature, partial melting of the mantle wedge cannot occur until a temperature of at least 1000°C is reached. Water released between ∼50 km and 100 km may remain within the slab or migrate into the adjacent mantle that is dragged downward (Figure 7). In either case, the upper part of the slab and mantle wedge have insufficient permeability to permit large-scale escape of volatiles, until partial melting of the mantle wedge occurs, with associated volatile advection. Of the volatiles that are released after much deeper subduction, some may trigger slab or adjacent mantle melting and reflux along the slab top to the arc-magma generation zone, and some may be lost to the deep mantle.

[104] Because partial melting in the mantle wedge is triggered by volatile release, a correlation between rate of arc volcanism and amount of deeply subducted water may exist. Early explorations of this hypothesis, using convergence rate as a proxy for subducted water, found no correlation between convergence rate and either arc eruption rate [Gill, 1981] or seismic-based arc growth rate [Reymer and Schubert, 1984]. Subducted water per unit arc length (Table 2) is well correlated (R = 0.72) with arc growth rate if arc ages of Jarrard [1986] are used to calculate the latter, but inversely correlated (R = −0.57) if arc ages from Reymer and Schubert [1984] are used; in both cases, 1–2 extreme points dominate the correlation. A comprehensive analysis of these hypotheses, using water fluxes from Table 2, may be warranted but is beyond the scope of this paper.

4.1.3. Slab Devolatilization and Benioff Zone Seismicity

[105] Devolatilization within the slab has been suggested as a mechanism for several aspects of Benioff zone seismicity. Metamorphic reactions and associated fluid release could control the updip and downdip limits of the large earthquakes generated at the contact between underthrusting and overriding plates [Hyndman et al., 1997; Kasahara et al., 2001]. Alternatively, the updip edge of interplate seismicity may be attributable to a suite of diagenetic changes that increase rock strength [Moore and Saffer, 2001], and the downdip edge may be compositionally determined, occurring at the overriding plate's Moho [Ruff and Tichelaar, 1996]. Dehydration is likely to elevate pore pressures and lower effective stress, thereby inducing earthquakes via hydrofracture. This dehydration embrittlement may be responsible for intermediate-depth earthquakes in subducting crust during transformation to eclogite [Kirby et al., 1996] and in the lower plane of double seismic zones during breakdown of serpentine to forsterite and enstatite [Peacock, 2001; Omori et al., 2001]. Serpentine dehydration may also be the cause of deep-focus earthquakes [Raleigh and Paterson, 1965; Meade and Jeanloz, 1991]. In all three cases, the wide depth distribution without a localized seismicity maximum is surprising.

[106] Slab devolatilization might also affect magnitude of interplate earthquakes: increased fluid pressure might decrease coupling between the plates, leading to smaller earthquakes. Rea and Ruff [1996] tested this hypothesis by comparing maximum earthquake magnitude to incoming sedimentary pore water for their 12 trench groups, and they found no correlation. They suggested, however, that a more diagnostic test would consider individual subduction zones rather than trench groups. For 23 subduction zones, incoming contents of bound and pore water of Table 2 can be compared to maximum earthquake magnitudes tabulated by McCaffrey [1997]. Water content is used rather than water flux, to remove the effects of trench length and convergence rate. This comparison confirms Rea and Ruff's [1996] conclusion that pore water content does not affect earthquake magnitude, and it extends this conclusion to include sediment bound water and total water from sediments and basalts. A weak correlation (R2 = 0.28 − 0.30) is found, but it is almost entirely attributable to a single subduction zone, Andaman, which has the highest content of incoming water and lowest earthquake magnitude. Convergence rate, rather than fluids, is the dominant control on earthquake magnitude, either because of coupling [Ruff and Kanamori, 1980; Jarrard, 1986] or recurrence time [McCaffrey, 1997]. Slab age may [Ruff and Kanamori, 1980; Jarrard, 1986] or may not [McCaffrey, 1997] be a secondary influence. Subducting fluids are not even a second-order influence, based on comparisons to earthquake magnitude residuals (observed minus predicted based on regression on convergence rate).

4.2. Chlorine Flux

[107] Assuming a typical seawater chlorinity of 19‰ (34.3‰ salinity) for subducted pore waters, 2.08 × 1013 g/yr of chlorine enters the world's subduction zones via pores in sediments and crust. Amounts of chlorine in subducting minerals are less reliably known but certainly much smaller. Ito et al. [1983] calculated that 2.9 ± 1.5 × 1012 g/yr of chlorine is subducted as oceanic crustal minerals, based on measured chlorine contents of the hydrous minerals and approximate mineral abundances; updated subduction rates would change this value to 2.5 ± 1.3 × 1012 g/yr. For chlorine within subducted sedimentary minerals, the geosynclinal average of 1200 ppm chlorine [Ronov and Yaroshevskiy, 1976] for a sedimentary subduction rate of 1.68 × 1015 g/yr provides a very rough estimate of 2 × 1012 g/yr. Total chlorine entering global subduction zones is therefore ∼2.5 × 1013 g/yr.

[108] Figure 8 presents a flowchart of global chloride fluxes. Although mantle degassing has transferred about 40% of Earth's chlorine to oceans and crust [Magenheim et al., 1995], the global chloride budget is probably near long-term mass balance, because salinity of the ocean has been relatively stable throughout at least the last 1.5 b.y. The oceanic chloride reservoir is so huge (2.7 × 1022 g), however, that even massive temporary imbalances have a negligible impact. The ocean loses chloride to cyclic sea salt and pores, and this loss is likely to be balanced by dissolved river influx and water reflux from the shallow subducted slab.


Figure 8. Flowchart of global chloride fluxes. See text for sources. All arc volcanism is shown as occurring on continents, but some occurs on oceanic crust that may later become continental crust.

Download figure to PowerPoint

[109] Chloride fluxes between the ocean and continents are 1–2 orders of magnitude larger than other chloride fluxes. Livingstone [1963] concluded that river influx of dissolved materials to the sea is 2.5–4 × 1015 g/yr, of which 6.5% is chloride, so the chloride influx is 1.6–2.6 × 1014 g/yr. Meybeck [1979] noted that pollution is often a major source of chloride in rivers, and he evaluated the natural river flux of chloride as 2.2 × 1014 g/yr. Much of this chloride, however, is “cyclic salt”, sea salt captured by the atmosphere and carried inland, where it is either deposited or washed out of the atmosphere by rain or snow [Clarke, 1924]. Direct measurement of the cyclic salt flux is not possible. Between 28% [Meybeck, 1983] and 82% [Berner and Berner, 1987] of the pollution-corrected chloride flux in rivers is attributable to dissolution of evaporites, and the rest is from cyclic salt. Consequently, cyclic sea salt transport of chloride from sea to land is ∼4–16 × 1013 g/yr.

[110] The mantle loses chloride to MORB and arc volcanism [Schilling et al., 1978] and to reflux, and it gains chloride from subduction. Using an average chlorine content of 48 ppm for MORB glasses, Ito et al. [1983] calculated a chloride supply rate of 2.7 × 1012 g/yr at spreading centers. Their estimation of chloride flux at arc volcanoes is subject to a factor of 20 uncertainty associated with magma generation rates [Ito et al., 1983]. Using an average chlorine content for arc magmas of 900 ± 500 ppm and the upper range of magma generation estimates (5–10 × 1015 g/yr), Ito et al. [1983] computed a chloride supply rate of 4.3–9.5 × 1012 g/yr. The MORB Cl flux is not changed by newer geochemical data [Jambon et al., 1995], nor by revised seafloor spreading rates and crustal thicknesses of this study. Revised arc magma generation (2.9–8.6 km3/yr) [Crisp, 1984] changes the arc Cl flux to 7–22 × 1012 g/yr.

[111] Reflux at and near the toes of accretionary prisms expels half of total subducted water prior to slab subduction to 15 km, and reflux may expel the two thirds of total subducted water released by 45 km (section 4.1.2). Associated chloride flux for the latter consists of 1.4 × 1013 g/yr from sediment pores and may include up to 0.6 × 1013 g/yr from basalts. The average chlorinity of reflux water is therefore 11–16‰, much less than the normal chlorinity of 19‰ for seawater and pore fluids. Chlorinity anomalies appear to be a common feature of décollements [Kastner et al., 1991]. For example, chlorinity of the Nankai décollement is ∼15‰, less than can be accounted for by smectite dehydration of subducting sediments and therefore implying an additional, deeper source of low-chlorinity water [Kastner et al., 1993; Underwood et al., 1993]. The Nankai chlorinity anomaly has been reproduced with a coupled flow and solute-transport model that incorporates transient flow within the sediments [Saffer and Bekins, 1998]. The water fluxes of Table 2 suggest that inclusion of smectite breakdown in the oceanic crust may be appropriate for reproducing the Nankai chlorinity anomaly; structural water in the Nankai upper crust is 50% higher than that in the subducted sediments.

[112] If the interpreted patterns of slab dewatering and associated reflux (Figure 7) are valid, then fluid inclusion studies of eclogites are predicted to exhibit the following chlorinities: ∼19‰ (seawater) for metasediments with possible smectite/illite transformation followed by compaction flushing, 6–13‰ (depending on crustal age) for metabasalts with partial expulsion of pore and structural water, and near-zero for metabasalts with extensive flushing from downdip portions of the slab. Metamorphic rocks from subduction zones (but not continental collisions) are consistently low-chlorinity [Bebout, 1996], generally 9–12‰ [Philippot et al., 1998].

[113] The mantle chloride budget may be close to mass balance currently (Figure 8), but uncertainties are large: total loss to arc and MORB volcanism and reflux is 2.4–4.5 × 1013 g/yr, and total gain from subduction is ∼2.5 × 1013 g/yr. In the near future, chlorine isotopic studies may be able to refine our understanding of chloride fluxes associated with both crustal alteration and subduction recycling [Magenheim et al., 1995].

4.3. CO2 Flux

[114] Global CO2 subduction totals 1.53 × 1014 g/yr. Two thirds of this amount is within the oceanic crust, particularly the upper portion of the extrusives, and the remainder is in accreted sediments (Figure 9). Nearly all of this CO2 is in the form of carbonate, primarily calcium carbonate. CO2 content of most subducting igneous crust varies within about a factor of two. In contrast, CO2 contents of subducting sediments are much more variable, with negligible carbonate and CO2 in half of the subducting sediment sections (Table 2). Variations in CO2 subduction fluxes among individual subduction zones (Figure 9) are most sensitive to convergence rate, trench length, and amount of subducting sedimentary carbonate.


Figure 9. Carbon dioxide fluxes at global subduction zones, from Table 2.

Download figure to PowerPoint

4.3.1. Comparison to Prior Flux Estimates

[115] My 5.26 × 1013 g/yr for CO2 in subducting sediments is based on very minor revisions of the comprehensive evaluation of Plank and Langmuir [1998]. These results supercede that of Peacock [1990], whose 1.6 × 1014 g/cc was based on a single hypothetical subducting sediment column that lacked hemipelagics and turbidites and was 72% carbonate; in contrast, Plank and Langmuir [1998] found that average subducting sediment is only 7% carbonate.

[116] Crustal CO2 subduction is calculated to be 1.00 × 1014 g/yr, within the wide range of previous estimates: 1.63 × 1014 [Staudigel et al., 1989], 0.6 × 1014 [Peacock, 1990], 1.8 × 1014 [Bebout, 1995, 1996], 1.18 × 1014 [Staudigel et al., 1996], and 1.48 × 1014 [Alt and Teagle, 1999]. These previous studies are based on estimating the CO2 content of typical old subducting crust. In contrast, flux calculations for the extrusives based on very young 6-Ma crust [Alt et al., 1996a] and 3.5 Ma hydrothermal fluids [Sansone et al., 1998] were an order of magnitude lower – 0.7–1.6 × 1013 and 0.4–1.1 × 1013, respectively.

[117] Peacock [1990] assumed 0.1% CO2 for both basalts and gabbros, based on consideration of whole-rock analyses from three upper-crustal sites and the gabbros of Hole 735B and Oman, whereas Bebout [1995, 1996] used 0.3% for the entire oceanic crust, based on data of Staudigel et al. [1989]. Staudigel et al. [1989, 1996] analyzed Holes 417A/417D/418A, and Alt and Teagle [1999] analyzed Sites 843 and 801. Both studies were carefully weighted to obtain a representative sample (see section 2.6 above) for the extrusives. Alt and Teagle [1999] extended this approach to encompass the dikes of Hole 504B and gabbros at three ODP localities. In contrast, Staudigel et al. [1989] interpolated between their value for the extrusive section and near-zero at 1100 m subbasement, and Staudigel et al. [1996] confined their flux assessment to the extrusive section.

[118] My CO2 flux calculation for oceanic crust is based largely on the excellent and comprehensive analysis of Alt and Teagle [1999]. Indeed, I adopt their CO2 values for the dikes and gabbros and for 4 of the 5 upper extrusive sections of Figure 2d. Rather than assume a uniform CO2 content for all altered crust, however, I use the tentative age-dependent CO2 enrichment pattern of Figure 2d to compute upper crustal CO2 for each subduction zone. The resulting total flux is only 70% of that of Alt and Teagle [1999], mainly because I use a total subduction rate that is only 71% of theirs (see section 3.1 discussion of revised subduction rates).

4.3.2. CO2 Flux Budget

[119] Consideration of the complete global CO2 flux budget (including mantle, continental crust, oceanic crust, oceans, and atmosphere) is beyond the scope of this paper. Berner et al. [1983] and François and Walker [1992], among others, presented CO2 flux budgets and models. My focus, in contrast, is on mantle and oceanic crust CO2 budgets.

[120] The dominant CO2 additions to the ocean/atmosphere system from crust and mantle sources are magma degassing at spreading centers and arc volcanoes, organic carbon breakdown, and possibly metamorphic breakdown of calcite to silicate. The dominant removal mechanisms are subaerial silicate weathering, which generates bicarbonate and culminates in calcite precipitation, carbonate uptake by oceanic crust, and organic carbon burial.

[121] Degassing at spreading centers contributes 2.8–5.5 × 1013 g/yr [Gerlach, 1989] or 5.7–13.6 × 1013 g/yr [Marty and Tolstikhin, 1998] of CO2; these values should probably be reduced ∼10% to account for lower spreading rates for NUVEL-1A than for earlier plate motion models. Degassing of arc volcanoes contributes another 11 ± 3 × 1013 g/yr [Zhang and Zindler, 1993; Marty and Tolstikhin, 1998] or 8–11 × 1013 g/yr [Kerrick, 2001], and plumes may add a comparable amount [Marty and Tolstikhin, 1998]. Subaerial weathering removes ∼2 × 1014 g/yr of CO2 from the atmosphere [Berner, 1990]. Burial of organic carbon removes 2.2 ± 0.9 × 1014 g/yr, but 9 ± 4 × 1013 g/yr is released by thermal breakdown [Berner, 1990]. Carbonate uptake by oceanic crust may approximately equal its current subduction rate, 1.0 × 1014 g/yr, or it may be much less (see sections 2.6 and 4.5). The ocean/atmosphere system must be close to steady state balance with crust and mantle, but that is not apparent from comparing these total CO2 atmosphere/ocean sources (3.5 ± 1.2 × 1014 g/yr) to sinks (5.6 ± 1.2 × 1014 g/yr). Additional possible sources, such as abiogenic methane and non-volcanic CO2 degassing, are discussed by Kerrick [2001]. The close coupling between CO2, climate, and silicate weathering may prevent major long-term imbalance between atmospheric CO2 sources and sinks [Berner and Caldeira, 2002].

[122] Subduction of sediments and oceanic crust removes ∼1.53 × 1014 g/yr of CO2 from the ocean/atmosphere system. In contrast to the initial subduction fluxes of water and chloride, most of which are lost to reflux, a negligible portion of subducted CO2 is lost to reflux. The CO2 content of reflux waters is likely to be close to saturation, or 0.02% for seawater at 0°C, so a water reflux of two-thirds initial subduction volume entrains ∼2.9 × 1011 g/yr of dissolved CO2, or only about 0.2% of subducting CO2.

[123] From the perspective of CO2 mass balance for the mantle, the subducting flux of CO2 exceeds the individual supplies from ridge volcanism and arc volcanism, but it is much less than the total degassing rate of ∼2.6 × 1014 g/yr (range 1.8–4.4 × 1014) [Marty and Tolstikhin, 1998]. Some crustal CO2 reaches the deep mantle, based on the excess of CO2 subduction over arc magmatism, the observation that CO2 appears to be mostly retained during metamorphism at depths of 15–45 km [Bebout, 1995], and the relatively large amount of CO2 still present in the mantle [Zhang and Zindler, 1993]. Phase equilibria suggest only minor decarbonation at and above subarc depths [Kerrick and Connolly, 2001], but dehydration enhances decarbonation and vice versa [Peacock, 1990; Kerrick and Connolly, 2001]. Kerrick and Connolly [1998] suggested that the CO2 in arc magmas is derived mainly from subducted sediments and extrusives, whereas CO2 within subducting mantle serpentinites (not considered in my flux calculations) is released in the deep mantle.

[124] The modern rate of crustal CO2 subduction is more significant than that of sedimentary CO2 for several reasons. First, total crustal CO2 subduction is double that from sediments. Second, all subducting crust contains significant CO2 whose ultimate release affects melting in the mantle wedge, whereas most sedimentary CO2 is in a few subduction zones. Third, the modern rate of crustal CO2 consumption is a robust indicator of long-term removal from the ocean/atmosphere system. In contrast, present-day sediment subduction is less relevant to ocean/atmosphere CO2 change than is carbon removal as deposition of calcium carbonate and organic matter. A significant portion of carbonate deposition [Hay, 1994] and most organic carbon deposition [Dymond and Lyle, 1994] are on continental shelves and are therefore isolated from mantle recycling for tens to hundreds of millions of years.

4.4. K2O Flux

[125] Sediments and igneous crust contribute about equal amounts to total subduction flux of K2O (Figure 10): 3.62 × 1013 g/yr and 3.81 × 1013 g/yr, respectively. Variations in K2O fluxes among individual subduction zones (Figure 10) are most sensitive to convergence rate, trench length, and thickness of subducting terrigenous sediments.


Figure 10. Potassium fluxes at global subduction zones, from Table 2.

Download figure to PowerPoint

4.4.1. Comparison to Prior Flux Estimates

[126] Subduction recycling of K2O was first quantified roughly by Sleep and Wolery [1979], who suggested that crustal alteration absorbs 1–2 × 1013 g/yr of K2O and that subducting red clays raise the total to ∼3 × 1013 g/yr. These values are less than half of mine. The estimated K2O flux of 3.62 × 1013 g/yr for subducting sediments is based on minor revisions to the comprehensive evaluation of Plank and Langmuir [1998]. Based on geochemical comparison of Holes 417A/417D/418A to unaltered crust, global K2O flux into the extrusives was estimated as 2.2 × 1013 g/yr [Spivack and Staudigel, 1994] and 1.95 × 1013 g/yr [Staudigel et al., 1996], similar to my 1.58 × 1013 g/yr for this zone. K2O fluxes have also been calculated for ophiolites [Lecuyer et al., 1990].

4.4.2. K2O Flux Budget

[127] The oceanic potassium budget (Figure 11) consists of supply from rivers (both dissolved and suspended load), sediment loss by deposition, diagenesis of seafloor sediments, K2O gain from oceanic crust during high-temperature alteration, and K2O loss to oceanic crust during low-temperature alteration. The mantle potassium budget, which is related to mantle thermal regime and 40Ar [Sleep, 1979; Sleep and Wolery, 1979; Jochum et al., 1983; Azbel and Tolstikhin, 1990], consists of loss to magmas during MORB, arc, and hot spot volcanism and plutonism, and gain during subduction of sediments and igneous oceanic crust. Neither budget need be in short-term or long-term balance. In particular, erosion roughly balances sediment subduction only on a time scale of tens to hundreds of millions of years.


Figure 11. Flowchart of global potassium fluxes. All arc volcanism is shown as occurring on continents, but some occurs on oceanic crust that may later become continental crust. Most suspended sediments transported to oceans are deposited on oceanic crust as shown, but some are deposited on continental margins.

Download figure to PowerPoint

[128] The oceanic K2O budget (Figure 11) is dominated by eroded K2O that is transported by rivers to the oceans as suspended sediment load and bedload, which total 3.7 × 1014 g/yr of K2O [Milliman and Meade, 1983]. Nearly all of this sediment is delivered to oceanic crust; only a few percent of Quaternary terrigenous marine sediments are deposited on continental shelves [Hay, 1994]. Annual river influx of dissolved K+ is also significant: based on a total dissolved load of 2.5–4 × 1015 g/yr, of which 1.9% is K+ [Livingstone, 1963], potassium transport (expressed as K2O) is 6–9 × 1013 g/yr; Meybeck [1979] estimated this influx as 6 × 1013 g/yr. Residence time of K2O in seawater (1.4 × 1024 g @ 0.38‰ K+) is therefore 8–13 m.y. A small portion of this dissolved potassium comes from airborne cyclic sea salt: 1–4 × 1012 g/yr (expressed as K2O) based on air transfer of chloride and the relative proportions of chloride and potassium in seawater. Potassium is absorbed from seawater by early diagenesis of aluminosilicates (zeolites and clays) at and very near the seafloor. Pore water K+ profiles indicate a global K+ flux of 1.1 × 1018 μEq/yr (5.2 × 1013 g K2O/yr) [Martin and Sayles, 1994]. Compared to most other oceanic potassium fluxes, the net removal of 5–7 × 1012 g/yr from high- and low-temperature crustal alteration is comparatively small. For example, I conclude that only 20% of the river input of potassium is absorbed during low-temperature alteration of oceanic crust (Figure 11), in contrast to an earlier 50% estimate [Spivack and Staudigel, 1994]. Calculated total input of dissolved K2O into the ocean (6.7 × 1013 g/yr; Figure 11) is surprisingly close to total output (6.6 × 1013 g/yr), considering the ∼10% errors in both.

[129] The subduction flux of crustal K2O includes both inherited magmatic K2O and alteration-induced changes. Earlier discussions of crustal K2O fluxes concentrated on addition of K2O to the extrusive section during off-axis low-temperature alteration. This process was recognized long ago to be one of the dominant geochemical signatures of low-temperature alteration [Hart et al., 1974; Donnelly et al., 1979b]. In contrast, high-temperature alteration of the dikes at and near the ridge axis removes K2O and is more deeply penetrating [Hart and Staudigel, 1982; Elderfield and Schultz, 1996].

[130] If the initial K2O content of basalts is 0.085% (section 2.8 and Figure 3), then only 22% of the 1.58 × 1013 g/yr extrusive K2O flux to subduction zones is primary. Both primary and alteration K2O fluxes for dikes are poorly known. Near-ridge high-temperature leaching apparently reduces the primary K2O dike flux (∼8.3 × 1012 g/yr) by 65–80%, to 1.4–2.9 × 1012 g/yr. The higher estimate assumes reduction from a typical MORB magmatic content of 0.085% K2O to the present Hole 504B content of 0.014%. The lower estimate assumes that the ∼65% K2O leaching of Hole 504B dikes, from an initial K2O of 0.04% [Shipboard Scientific Party, 1992], is representative of dikes in general, but leaching is even less if initial K2O is less. The 2.15 × 1013 g/yr K2O in subducting gabbros appears to be almost entirely primary; average K2O at Hole 735B is 0.05% for fresh rocks [Bach et al., 2001] and 0.06% for all rocks (section 3.7). Thus, only ∼13% of subducted crustal K2O is attributable to addition from seawater; the rest (3.31 × 1013 g/yr) is inherited from MORB magmas. Because more than half of the total crustal K2O subducted is in the gabbros, any alteration-induced change in gabbro K2O could easily surpass those from the extrusives and dikes.

[131] The K2O budget for igneous oceanic crust (excluding sediments) almost balances, with 4.9 × 1013 g/yr input and 4.3–4.5 × 1013 g/yr output (Figure 11). The slight imbalance is because the rate of crustal generation is ∼10% higher than that of subduction: ∼10% of convergence is continent/continent collision, so total K2O flux from mantle to oceanic crust at spreading centers is 3.64 × 1013 g/yr (= 1.1 × 3.31 × 1013 g/yr).

[132] The largest uncertainty in global K2O fluxes is in the dominant component, K2O flux from the mantle to arc magmatism, mainly because of uncertainties in the ratio of observed arc volcanism to hidden arc plutonism. Based on 0.4–0.6 km3/yr of arc volcanism and 2.5–8 km3/yr of arc plutonism [Crisp, 1984], a density of 2.7 g/cm3, and average K2O of 3.8% for silicic igneous rock and 5.5% for calc-alkalic granite [Nockolds, 1954], the total K2O flux due to arc magmatism is 4–12 × 1014 g/yr. This flux to continents is comparable to the total K2O flux from continents to oceans of 4.3–4.7 × 1014 g/yr, but it is more than an order of magnitude higher than the MORB crustal K2O flux and about an order of magnitude higher than the combined subduction flux of 7.4 × 1013 g/yr from sediments and oceanic crust.

[133] K2O fluxes from subduction and arc magmatism need not balance: most of the latter is thought to come from low-degree partial melting of the mantle wedge rather than the slab [e.g., Ionov et al., 1997], as required by the flux imbalance of Figure 11. Arc K2O is accompanied by long-term mantle 40Ar, and helium similarity of arc and MORB volcanism suggests similar source regions [Sleep and Wolery, 1979]. In contrast, the systematic increase in arc magma K2O with increasing slab depth [Dickinson and Hatherton, 1967; Dickinson, 1975] suggests a slab contribution. K2O is scoured from the slab and released through solution in fluids, and amount of K2O release may depend more on dehydration history than on initial concentrations of K2O or hydrous minerals [Leeman, 1996]. Metabasalts from <45 km burial depth retain their pre-subduction K2O concentrations [Bebout, 1995]. Most of the hydrous phases that break down within the slab at depths of <110 km are potassium-free. In contrast, phengite (a K-rich mica) could be the dominant K-rich hydrous phase to break down at depths of >100 km and enrich arc magmas [Schmidt, 1996].

[134] Subducted water has been suggested to affect arc magma chemistry, based on comparison of arc K2O contents to convergence rates [Sugisaki, 1976], but confounding variables such as continental contamination and degree of partial melting were not considered. Plank and Langmuir [1993] minimized these confounding effects by considering only primitive arc basalts and normalizing arc potassium contents to sodium. For eight analyzed subduction zones, they found a very good correlation (R2 = 0.739) between arc potassium content and the amount of potassium subducted in sediments. This correlation is just as good with the parameters of Table 2 (R2 = 0.763), but it is significantly degraded if crustal K2O sources are included: R2 = 0.624 including upper extrusives, and R2 = 0.044 for total subducted K2O. This pattern suggests that K2O recycling to arc magmas is largely confined to subducted sediments.

[135] The present-day imbalance between K2O subduction (7.5 × 1013 g/yr) and arc magmatism (∼4–12 × 1014 g/yr) is expected, based on the overall imbalance between 1.60 × 1015 g/yr of terrigenous sediment subduction [Plank and Langmuir, 1998] and 10–20 × 1015 g/yr of terrigenous sediment deposition [Holland, 1981; Milliman and Syvitski, 1994]. Rea and Ruff [1996] explored the implications of the terrigenous imbalance and concluded that mass balance is likely to prevail only over the hundreds-of-million-years time scale of Wilson cycles. Alternatively, long-term growth of continents may result; Reymer and Schubert [1984] used volumes and ages of arcs to estimate that this growth averages ∼4 × 1015 g/yr.

4.5. Are Present Fluxes Representative of Prior Ones?

[136] Modern subduction fluxes of H2O, CO2, K2O, and Cl are probably quite different from those that characterized early earth history. Different modeling approaches agree that most volatile release from the mantle occurred quite early [McGovern and Schubert, 1989; Franck and Bounama, 1997; Harrison, 1999]. Continental growth was probably also mainly Archean [e.g., Harrison, 1999], but rate of Phanerozoic growth is debated [Reymer and Schubert, 1984]. Thermal modeling suggests that Precambrian spreading rates were no more than twice modern ones [Sleep, 1979]. Rapid Archean spreading implies generally younger slabs, which may have dehydrated at a shallower depth than modern ones [McCulloch, 1993], but rapid subduction also implies retarded slab heating, which leads to deeper devolatilization [Staudigel and King, 1992; Kerrick and Connolly, 2001]; the latter is likely to dominate. Archean ophiolites of South Africa and Australia appear to have undergone hydrothermal histories quite similar to modern oceanic crust [DeWit and Hart, 1993]. Phanerozoic variations in seafloor spreading rate [Gaffin, 1987] have associated changes in MORB degassing, subduction fluxes, and possibly arc degassing [François and Walker, 1992].

[137] Modern subduction fluxes of H2O, CO2, K2O, and Cl are sensitive to the volume of subducting sediments, which changes dramatically as plate geometries evolve. Rea and Ruff [1996] discussed this problem in the context of the order-of-magnitude discrepancy between current terrigenous sedimentation and subduction. von Huene and Scholl [1991] appraised the long-term rate of sediment subduction as only about two thirds of the current rate, because sedimentation rates have been unusually rapid during the last 15–30 m.y. [Rea, 1993]. These rates have been even more rapid during the Quaternary, associated with increased glacial activity [Hay, 1994]; this factor affects sediment subduction calculations incorporating trench fill [von Huene and Scholl, 1991] much more than those based on distal reference sites [Rea and Ruff, 1996; Plank and Langmuir, 1998; this study].

[138] Comparisons of subduction fluxes to the dissolved and suspended loads of rivers (e.g., Figures 8 and 11) are severely limited by temporal changes in the latter: river fluxes for the Recent are nonrepresentative of earlier fluxes [Hay, 1994; Meybeck, 1994], and both dissolved and suspended loads have varied by one order of magnitude in the past [Meybeck, 1994]. For example, the suspended-sediment estimate of Milliman and Meade [1983], which we used for K2O, has been revised upward by 50% [Milliman and Syvitski, 1994], but the flux prior to widespread farming was less than half of this value [Milliman and Syvitski, 1994].

[139] Late Mesozoic and Cenozoic variations in atmospheric CO2 have probably affected the observed pattern of age-dependent upper crustal CO2 (section 2.6). More rapid seafloor spreading rates during the Cretaceous [Kominz, 1984; Larson, 1991] have been causally linked not only to higher sea levels [Hays and Pitman, 1973], but also to higher atmospheric CO2 levels and consequently warmer temperatures [Berner et al., 1983; François and Walker, 1992]. Rate of ridge-axis CO2 degassing is proportional to spreading rate, but this effect is counteracted by increased subduction of CO2-rich sediments and crust; the net effect is flux to the atmosphere/ocean system if generally younger, less altered crust with thinner sediments is subducted. Faster subduction may foster increased arc CO2 degassing, but this assumption is hazardous [Kerrick, 2001]. Furthermore, calcareous phytoplankton deposition has fluctuated widely [Nakamori, 2001]. Numerical modeling of the Phanerozoic carbon cycle, attempting to consider most of these variables and their feedbacks, suggests that the dominant temporal change is either seafloor alteration [François and Walker, 1992] or–more likely–terrestrial silicate weathering [Caldeira, 1995]. Alternatively, individual CO2 enrichment events may have different sources [Kerrick, 2001].

[140] The “modern” subduction fluxes of this study are actually time-averaged fluxes, representative of several million years. As mentioned previously (section 4.1.2), present-day fluid fluxes from accretionary prisms, observed at the seafloor via submersibles or inferred from heat flow, are much higher than long-term averages [e.g., Le Pichon et al., 1991; Kastner et al., 1991]. This discrepancy is significantly reduced by considering not just shallow sediment sources but also deeper reflux water from both crust and sediments. For Nankai, Peru, and Barbados, whose long-term fluid fluxes were estimated as 7 m3/yr per m along trench [Kastner et al., 1991], revised fluxes are 17, 24, and 35 m3/yr per m (not including compaction within each prism), based on Table 2 and the conclusion that two thirds of subducted water is refluxed. Transient flow can account for the original discrepancies [Moore and Vrolijk, 1992; Le Pichon et al., 1993], with a flux cycle on an individual trench lasting perhaps ∼2.5–5 Ma [Saffer and Bekins, 1998]. If most of a several-million-year budget of subducted water and CO2 of a high-flux subduction zone like Makran were refluxed within a few-thousand-year period, it would raise sea level by ∼1 m and double atmospheric CO2. However, the sea level effect is minor in comparison to ice-induced sea level change, and CO2 reflux is unlikely to exceed saturation levels.

4.6. Uncertainties

[141] A weakness of the preceding flux analyses is the inability to attach confidence limits to computed fluxes. Calculation of confidence limits from regression relations such as Figure 2 is straightforward, as is formal uncertainty analysis for means given standard deviation and number of points. The problem, however, is that systematic errors are likely to overwhelm random errors in most cases. For example, mean microporosity for gabbros of Hole 735B is 0.74%, with 95% confidence limits of ±0.29%, but this value is probably strongly biased by unroofing-induced fracturing; in situ microporosities of most gabbros might be much lower. This section focuses on some of the dominant uncertainties - particularly biases - that affect these flux estimates. An Excel spreadsheet combining the calculations of Tables 1 and 2 is included as an auxiliary material file, to encourage readers to supercede my calculations by revising parameters or assumptions.

4.6.1. Unquantified Variables Affecting Upper Crustal Alteration

[142] The assumed alteration mineralogy for upper oceanic crust–saponite, celadonite, and calcite (section 2.8)–is only a first-order approximation. Actual alteration mineralogy is invariably more complex, but this alteration model probably provides a reasonable indication of the age-dependent enrichment of K2O and H2O+ within the upper extrusives.

[143] The time-dependent alteration of upper oceanic crust, as well as associated changes in geophysical properties, can be locally overwhelmed by four other sources of variance.

[144] 1. Initial permeability is sensitive to porosity and therefore also to volcanic style. Flows are generally much more massive and impermeable than pillows, so they are less subject to alteration. Consequently, systematic age-dependent changes in log-scale velocity, core-plug velocity, and macroporosity are much stronger for pillows (significant at 95%, 95%, and 99% c.l., respectively) than for flows (none is significant) [Jarrard et al., 2003]. Volcanic style is the dominant variable controlling within-site alteration variations, but the ubiquitous preponderance of pillows over flows prevents volcanic style from obscuring the age-dependent relationships.

[145] 2. Local basement topography often controls the geometry of hydrothermal circulation, concentrating outflow at topographic highs [Fisher et al., 1990, 1994]. This pattern is evident for paired holes on adjacent topographic highs and saddles, at 417A/417D/418A and at 504B/896A.

[146] 3. Spreading rate affects crustal structure (see section 3.3) and therefore 3-D permeability structure. A relationship between spreading rate and crustal alteration is likely, but none has been detected yet.

[147] 4. Sedimentation restricts access of seawater to crustal porosity. Crustal age is more globally important than sedimentation in determining the transition from convective to conductive heat loss [Stein and Stein, 1994], but sedimentation can locally control hydrothermal circulation and resulting alteration. These analyses exclude areas such as the Juan de Fuca sedimented ridge where sedimentation undoubtedly dominates, yet they include Hole 504B which has sufficient sediment blanket for more conductive than convective heat flow [Fisher et al., 1990] despite its young age.

[148] These four sources of local variation in alteration are probably the main causes of dispersion among sites in Figure 2, but they do not introduce systematic bias into these patterns. Fortunately, statistically significant age-dependent crustal alteration patterns are evident despite this variance: macroporosity reduction and consequent chloride expulsion (99.9% confidence level), CO2 enrichment (97.5% c.l.), matrix density reduction and consequent hydration (99.9% c.l.), and K2O enrichment (based on age versus matrix density @99.9% c.l. and matrix density versus K2O @99% c.l.).

[149] Calculated global fluxes are relatively insensitive to the assumed patterns of age-dependent upper crustal alteration. If age-dependence is ignored and the upper crustal alteration state of all subduction zones is assumed to be that of 77-Ma crust, the global average age [Sprague and Pollack, 1980], global fluxes change by 12% for CO2, 3% for K2O, 2% for bound water, and −3% for pore water and chloride. The impact on calculated fluxes for individual subduction zones is much higher, ranging from −19% to +28% for most fluxes and >80% for CO2 in young subducting crust.

4.6.2. Rejuvenation of Hydrothermal Circulation

[150] The geochemical signature of age-dependent crustal alteration, though only roughly known, can provide an essential starting point for geochemical mass balances associated with subduction. Implicit to such calculations, however, is the assumption that normal oceanic crust is not modified by the subduction process itself. This assumption may be invalid.

[151] The flexure of lithosphere prior to subduction generates an outer rise, with associated normal faulting that might rejuvenate hydrothermal circulation. The two lowest matrix densities on Figure 2b are Holes 1149B and 1149D, the only holes that have sampled basaltic basement from an outer rise [Shipboard Scientific Party, 2000b]. If this correlation is not coincidental, analyses of geochemical fluxes at subduction zones should incorporate the effects of hydrothermal rejuvenation within oceanic crust immediately seaward of the trench axis. The effects of extensional cracking extend to depths of at least 25 km [Christensen and Ruff, 1988], much deeper than the crust. These cracks are likely to permit fluid percolation into the mantle and induce serpentinization [Peacock, 2001].

4.6.3. Properties of Gabbros

[152] The properties assumed for gabbros (section 3.7 and Table 1) are based on Hole 735B, with supporting evidence from Site 894, but both sites have been affected by the unroofing that made gabbro drilling possible. For example, I introduced section 4.6 with the example of gabbro microporosity bias. Furthermore, Hole 735B appears to be nonrepresentative of the geochemically heterogeneous lower crust [Dick et al., 2000]. Gabbro uncertainties have a disproportionate impact on mass flux calculations, because 70% of oceanic crust is gabbro. The impact is greatest for K2O fluxes: section 3.7 noted that determining gabbro K2O based on the K2O log rather than XRF would raise the global subduction flux for crustal K2O by 71%. The high-temperature K2O fluxes between gabbros and seawater, if any, are undetermined.

4.6.4. Peridotite Subduction Fluxes

[153] This study considers only those subduction fluxes associated with oceanic crust and its overlying sediments. Mantle peridotites are excluded, although they comprise the majority of subducting lithosphere, because their in situ alteration state is virtually unknown [e.g., Peacock, 1996]; obtainable samples are unlikely to be representative of normal lower lithosphere. Peacock and Hyndman [2001] noted that typical upper mantle velocities of 8.2 km/s are consistent with anhydrous peridotite. Dehydration of serpentinites, which contain 13% H2O+, could be a major source of the H2O in arc magmas [Ulmer and Trommsdorff, 1995; Kerrick and Connolly, 1998], and subducting ophicarbonates (calcite-rich serpentinites) may transport CO2 to the deep mantle [Kerrick and Connolly, 1998]. Even 5% serpentinization of the upper mantle would constitute about one third of the subducting water flux [Peacock and Hyndman, 2001].

4.6.5. Subduction Erosion and Underplating

[154] The comparisons of subduction fluxes to arc fluxes in this paper neglect the possibility that the slab may exchange material with the overriding plate through subduction erosion or underplating. The inverse correlation between shallow slab dip and arc age (duration of subduction) [Jarrard, 1986] implies substantial subduction erosion, and von Huene and Scholl [1991] estimated that this erosional flux may be comparable to the flux of sediment subduction. Both subduction erosion and underplating are poorly known [Dewey and Windley, 1981]; neglecting them, however, biases comparisons of subduction fluxes to arc fluxes.

5. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[155] High-temperature alteration of the oceanic crust occurs at and very near the spreading center; reference Holes 504B and 735B offer perspectives on this alteration for the dikes and gabbros, respectively. In contrast, a dozen logged DSDP and ODP sites document the progressive, age-dependent low-temperature alteration of extrusives. Two main variables control alteration of the extrusives: (1) extrusive type (flows, pillows, hyaloclastites), which affects permeability and therefore water/rock ratio, and (2) time. Intersite variations in proportions of flows versus pillows are a principal source of local variance: pillows are much more porous, permeable, and altered than more massive flows. Regionally, however, this variance is averaged, and the waning impact of off-axis hydrothermal circulation on geophysical properties is evident. Seismic velocity and permeability detect only the large changes during the first 10 Ma after crustal formation [Carlson, 1998; Fisher and Becker, 2000], but some properties have the resolving power to detect subtler off-axis changes. I identified the following systematic changes in properties of the upper extrusives, each of which appears to wane with a log(age) dependence, throughout the lifetime of oceanic crust: (1) macroporosity reduction, (2) intergranular alteration to smectite, celadonite, and calcite, with the calcite cementation occurring generally later than the others, and (3) increase in large-scale velocity due to macroporosity reduction, mainly in pillows. These changes involve progressive uptake of H2O+, CO2, and K2O, with attendant expulsion of pore water and its Cl. Given crustal age, one can calculate the average concentrations of pore H2O, H2O+, CO2, Cl, and K2O in the extrusives.

[156] Surprisingly, low-temperature alteration causes no net change in total water. Macroporosity is replaced with alteration minerals containing some structural water, and intergranular alteration forms hydrous minerals, but pore water loss is nearly identical to H2O+ gain. Chloride is presumably expelled by these transformations, but reliable pore water samples have not been obtained from the extrusives. The overall alteration effect on total crustal content of K2O is also smaller than expected. Enrichment by low-temperature alteration of the upper extrusives is obvious at abundant DSDP and ODP crustal sites, but substantial K2O depletion occurs within the dikes during high-temperature alteration. The majority of crustal K2O for all crustal ages is primary rather than secondary, and most is in the gabbros. The largest net geochemical flux from crustal alteration, by an order of magnitude, is the CO2 enrichment that occurs at both intergranular scale and via fracture filling. Its pattern of age-dependence is yet to be documented in detail, but existing data are compatible with the log(age) dependence seen for crustal geophysical properties (Figure 2).

[157] By applying these predictions to 41 modern subduction zones, one can determine modern mass fluxes of pore H2O, H2O+, CO2, Cl, and K2O, both for individual subduction zones and globally. This data set is complemented by flux determinations for the same components within subducting sediments at 26 of these subduction zones by Plank and Langmuir [1998].

[158] Subducted crustal water is released and expelled at various slab depths by several mechanisms, depending on how the water is held–in sediment pores, in sedimentary smectites, in basalt pores, in loosely bound hydrous minerals in the extrusives, and in relatively stable hydrous minerals within the dikes and gabbros (Figure 7). At burial depths of <10 km, sediment compaction dominates, with additional water coming from smectite/illite conversion in both sediments and extrusive basalts. At greater depths, metamorphic reactions convert pore water to structural and expelled waters. I estimate that half to two thirds of subducted water is later refluxed at or near the toe of the prism, rising along the slab from at least 15 km and possibly as deep as 45 km. This long-distance transport is inferred from patterns of dehydration along slab-parallel permeability at 15–45 km in the Catalina Schist [Bebout, 1995, 1996] and from magnetotelluric evidence of free water in slabs at 15–60 km depth. Crustal dehydration may be as important as transformation of sedimentary smectite to illite in generating the low-chlorinity anomalies seen in some décollements.

[159] The remaining water is more deeply subducted as stable hydrous minerals. Much or most of this water is released by a series of dehydration reactions at depths of 45–100 km. It is carried along with the slab until it escapes by triggering the partial melting that permits it to reach Earth's surface at the arc. Surprisingly, there is little evidence that amount of subducted volatiles affects either amount of magma generation or earthquake magnitude on the interplate boundary, despite indications that reflux water traverses the entire boundary.

[160] Global flux balances for H2O, CO2, Cl, and K2O are constrained partly by the ocean/crust interchanges documented above, partly by the subduction fluxes above, and additionally by other published data. In general, fluxes among oceans, oceanic crust, continental crust, and mantle balance, within the uncertainties. A major exception is the present order-of-magnitude excess of marine sedimentation over sediment subduction, most obvious in the K2O and CO2 budgets but also affecting porosity and therefore H2O and Cl. Presumably, a closer balance is achieved on a time scale of hundreds of millions of years, though some imbalance may persist in the form of long-term growth of continents. Temporal variation in dissolution and precipitation of evaporites can unbalance the global chloride budget dramatically, without significantly affecting ocean salinity.

[161] A brief overview of CO2 mass fluxes confirms that carbonate precipitation in the oceanic crust is a major CO2 sink, but it provides an unsatisfactory balance of mantle and crust input versus output. Subducted CO2 reaches subarc depths with minimal loss to reflux, supplying the CO2 erupted in arc magmas and replenishing deeper-mantle CO2.

[162] The oceanic potassium budget is well balanced, with almost as much loss to seafloor diagenesis as gain from dissolved river input. Some signature of subducted sediment K2O on arc magma K2O has been demonstrated [Plank and Langmuir, 1993]. The major excess of arc magma K2O over slab K2O, however, requires that most of the former comes from the mantle wedge and crustal enrichment.


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[163] I am grateful to J. C. Alt and an anonymous reviewer for their constructive reviews. This project was made possible by the data acquisition efforts of shipboard scientific parties, operations managers, and particularly the logging scientists of two dozen DSDP and ODP legs. Just as important to the success of this project was the availability of log and core data, on CD-ROM and at the Web sites of ODP Logging Services and JANUS. This research used samples and data provided by DSDP and ODP. ODP is sponsored by the U.S. National Science Foundation and participating countries under management of Joint Oceanographic Institutions, Inc.


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Alt, J. C., Subseafloor processes in mid-ocean ridge hydrothermal systems, in Seafloor Hydrothermal Systems, Physical, Chemical, Biological, and Geological Interactions, Geophys. Monogr. Ser., vol. 91, edited by S. E. Humphris et al., pp. 85114, AGU, Washington, D.C., 1995.
  • Alt, J. C., Very low-grade hydrothermal metamorphism of basic igneous rocks, in Low-Grade Metamorphism, edited by M. Frey, and D. Robinson, pp. 85114, Blackwell Sci., Malden, Mass., 1999.
  • Alt, J. C., and J. Honnorez, Alteration of the upper oceanic crust, Deep Sea Drilling Project Site 417: Mineralogy and chemistry, Contrib. Mineral. Petrol., 87, 149169, 1984.
  • Alt, J. C., and D. A. H. Teagle, The uptake of carbon during alteration of ocean crust, Geochim. Cosmochim. Acta, 63, 15271535, 1999.
  • Alt, J. C., and D. A. H. Teagle, Hydrothermal alteration and fluid fluxes in ophiolites and ocean crust, in Ophiolites and Oceanic Crust: New Insights from Field Studies and the Ocean Drilling Program, edited by Y. Dilek et al., Geol. Soc. Am. Spec. Pap., 349, 273282, 2000.
  • Alt, J. C., C. Laverne, and K. Muehlenbachs, Alteration of the upper oceanic crust: Mineralogy and processes in DSDP Hole 504B, Leg 83, Initial Rep. Deep Sea Drill. Proj., 83, 217247, 1985.
  • Alt, J. C., J. Honnorez, C. Laverne, and R. Emmermann, Hydrothermal alteration of a 1-km section through the upper oceanic crust, Deep Sea Drilling Project Hole 504B: Mineralogy, chemistry and evolution of seawater-basalt interactions, J. Geophys. Res., 91, 10,30910,335, 1986.
  • Alt, J. C., C. France-Lanord, P. A. Floyd, P. Castillo, and A. Galy, Low-temperature hydrothermal alteration of Jurassic ocean crust, Site 801, Proc. Ocean Drill. Program Sci. Results, 129, 415427, 1992.
  • Alt, J. C., D. A. H. Teagle, C. Laverne, D. Vanko, W. Bach, J. Honnorez, K. Becker, M. Ayadi, and P. A. Pezard, Ridge flank alteration of upper ocean crust in the eastern Pacific: A synthesis of results for volcanic rocks of Holes 504B and 896A, Proc. Ocean Drill. Program Sci. Results, 148, 435452, 1996a.
  • Alt, J. C., et al., Hydrothermal alteration of a section of upper oceanic crust in the eastern equatorial Pacific: a synthesis of results from Site 504 (DSDP Legs 69, 70, and 83, and ODP Legs 111, 137, 140, and 148), Proc. Ocean Drill. Program Sci. Results, 148, 417434, 1996b.
  • Anderson, R. N., and M. D. Zoback, Permeability, underpressures and convection in the oceanic crust near the Costa Rica Rift, eastern equatorial Pacific, J. Geophys. Res., 87, 28602868, 1982.
  • Anderson, R. N., S. Uyeda, and A. Miyashiro, Geophysical and geochemical constraints at converging plate boundaries; Part I. dehydration in the downgoing slab, Geophys. J. R. Astron. Soc., 44, 333357, 1976.
  • Anderson, R. N., M. G. Langseth, and J. G. Sclater, The mechanisms of heat transfer through the floor of the Indian Ocean, J. Geophys. Res., 82, 33913409, 1977.
  • Anderson, R. N., S. E. DeLong, and W. M. Schwarz, Dehydration, asthenospheric convection, and seismicity in subduction zones, J. Geol., 88, 445451, 1980.
  • Andrews, A. J., Saponite and celadonite in layer 2 basalts, DSDP Leg 37, Contrib. Mineral. Petrol., 73, 323340, 1980.
  • Armstrong, R. L., A model for the evolution of strontium and lead isotopes in a dynamic Earth, Rev. Geophys., 6, 175199, 1968.
  • Armstrong, R. L., The persistent myth of crustal growth, Aust. J. Earth Sci., 38, 613630, 1991.
  • Arzate, A. J., M. Mareschal, and D. Livelybrooks, Electrical image of the subducting Cocos Plate from magnetotelluric observations, Geology, 23, 703706, 1995.
  • Azbel, I. Y., and I. N. Tolstikhin, Geodynamics, magmatism, and degassing of the Earth, Geochim. Cosmochim. Acta, 54, 139154, 1990.
  • Bach, W., J. C. Alt, Y. Niu, S. E. Humphris, J. Erzinger, and H. J. B. Dick, The geochemical consequences of late-stage low-grade alteration of lower ocean crust at the SW Indian Ridge: Results from ODP Hole 735B (Leg 176), Geochim. Cosmochim. Acta, 65, 32673287, 2001.
  • Bach, W., S. E. Humphris, and A. T. Fisher, Fluid flow and fluid-rock interaction within ocean crust: reconciling geochemical, geological, and geophysical observations, in Subsurface Biosphere at Mid-Ocean Ridges, Geophys. Monogr. Ser., edited by W. Wilcock et al., AGU, Washington, D.C., in press, 2003.
  • Bebout, G. E., Geometry and mechanisms of fluid flow at 15 to 45 kilometer depths in an Early Cretaceous accretionary complex, Geophys. Res. Lett., 18, 923926, 1991.
  • Bebout, G. E., The impact of subduction-zone metamorphism on mantle-ocean chemical cycling, Chem. Geol., 126, 191218, 1995.
  • Bebout, G. E., Volatile transfer and recycling at convergent margins: Mass-balance and insights from high-P/T metamorphic rocks, in Subduction Top to Bottom, Geophys. Mon., vol. 96, edited by G. E. Bebout et al., pp. 179193, AGU, Washington, D.C., 1996.
  • Becker, K., et al., Drilling deep into young oceanic crust, Hole 504B, Costa Rica Rift, Rev. Geophys., 27, 79102, 1990.
  • Beiersdorf, H., et al., Age and possible modes of formation of the Celebes Sea basement, and thermal regimes within the accretionary complexes off SW Mindanao and N Sulawesi, paper presented at International Conference on Stratigraphy and Tectonic Evolution of Southeast Asia and the South Pacific, GEOTHAI'97 conference, Int. Union of Geol. Sci., Bangkok, Thailand, 1997.
  • Bell, D. R., and G. R. Rossman, Water in Earth's mantle: The role of nominally anhydrous minerals, Science, 255, 13911397, 1992.
  • Berner, E. K., and R. A. Berner, The Global Water Cycle: Geochemistry and Environment, 397 pp., Prentice-Hall, Old Tappan, N.J., 1987.
  • Berner, R. A., Global CO2 degassing and the carbon cycle: Comment on “Cretaceous ocean crust at DSDP sites 417 and 418: Carbon uptake from weathering versus loss by magmatic outgassing,” Geochim. Cosmochim. Acta, 54, 28892890, 1990.
  • Berner, R. A., and K. Caldeira, The geologic carbon cycle and the evolution of atmospheric carbon dioxide, Eos Trans. AGU, 83(47), Fall Meet. Suppl., F383, 2002.
  • Berner, R. A., and Z. Kothavala, GEOCARB III: A revised model of atmospheric CO2 over Phanerozoic time, Am. J. Sci., 301, 182204, 2001.
  • Berner, R. A., A. C. Lasaga, and R. M. Garrels, The carbonate-silicate geochemical cycle and its effect on atmospheric carbon dioxide over the past 100 million years, Am. J. Sci., 283, 641683, 1983.
  • Bevis, M., et al., Geodetic observations of very rapid convergence and back-arc extension at the Tonga arc, Nature, 374, 249251, 1995.
  • Booij, E., and H. Staudigel, Controls and diversity of altered oceanic crust composition, Eos Trans. AGU, 78(46), AGU Fall Meet. Suppl., F806, 1997.
  • Bourbié, T., O. Coussy, and B. Zinszner, Acoustics of Porous Media, 334 pp., Educ. Technol., Paris, 1987.
  • Bray, C. J., and D. E. Karig, Porosity of sediments in accretionary prisms and some implications for dewatering processes, J. Geophys. Res., 90, 768778, 1985.
  • Brias, A., P. Patriat, and P. Tapponnier, Updated interpretation of magnetic anomalies and seafloor spreading stages in the South China Sea: Implications for the Tertiary tectonics of Southeast Asia, J. Geophys. Res., 98, 62996328, 1993.
  • Broglia, C., and D. Ellis, Effect of alteration, formation absorption, and standoff on the response of the thermal neutron porosity log in gabbros and basalts: Examples from Deep Sea Drilling Project-Ocean Drilling Program sites, J. Geophys. Res., 95, 91719188, 1990.
  • Brown, K. M., and B. Ransom, Porosity corrections for smectite-rich sediments: Impact on studies of compaction, fluid generation, and tectonic history, Geology, 24, 843846, 1996.
  • Buckley, H. A., J. C. Bevan, K. M. Brown, J. R. Johnson, and V. C. Farmer, Glauconite and celadonite: two separate mineral species, Mineral. Mag., 42, 373382, 1978.
  • Busch, W. H., P. R. Castillo, P. A. Floyd, and G. Cameron, Effects of alteration on physical properties of basalts from the Pigafetta and East Mariana Basins, Proc. Ocean Drill. Program Sci. Results, 129, 485499, 1992.
  • Caldeira, K., Long-term control of atmospheric carbon dioxide: Low-temperature seafloor alteration or terrestrial silicate-rock weathering? Am. J. Sci., 295, 10771114, 1995.
  • Cannat, M., How thick is the magmatic crust at slow-spreading oceanic ridges? J. Geophys. Res., 101, 28472857, 1996.
  • Carlson, R. L., Seismic velocities in the uppermost oceanic crust: Age dependence and the fate of layer 2A, J. Geophys. Res., 103, 70697077, 1998.
  • Carlson, R. L., Lower crustal water contents, P-wave velocities, and modal mineralogy of oceanic diabase and gabbro, Eos Trans. AGU, 82(47), Fall. Meet. Suppl., F1154, 2001.
  • Christensen, D. H., and L. J. Ruff, Seismic coupling and outer rise earthquakes, J. Geophys. Res., 93, 13,42113,444, 1988.
  • Christensen, N. I., and M. H. Salisbury, Velocities, elastic moduli and weathering-age relations for Pacific Layer 2 basalts, Earth Planet. Sci. Lett., 19, 461470, 1973.
  • Christensen, N. I., and M. H. Salisbury, Seismic velocities, densities and porosities of Layer 2B and Layer 2C basalts from Hole 504B, Initial Rep. Deep Sea Drill. Proj., 83, 367370, 1985.
  • Christensen, N. I., and J. D. Smewing, Geology and seismic structure of the northern section of the Oman Ophiolite, J. Geophys. Res., 86, 25452555, 1981.
  • Christensen, N. I., W. W. Wepfer, and R. D. Baud, Seismic properties of sheeted dikes from Hole 504B, ODP Leg 111, Proc. Ocean Drill. Program Sci. Results, 111, 171176, 1989.
  • Clarke, W. M., The Data of Geochemistry, U. S. Geol. Surv. Bull., 770, 841 pp., 1924.
  • Coleman, R. C., Ophiolites: Ancient Oceanic Lithosphere?, 229 pp., Springer-Verlag, New York, 1977.
  • COSOD-II, Report of the Second Conference on Scientific Ocean Drilling, Strasbourg, 6–8 July, 1987, 142 pp., JOIDES, Washington, DC, 1987.
  • Crisp, J. A., Rates of magma emplacement and volcanic output, J. Volcanol. Geotherm. Res., 20, 177211, 1984.
  • Davis, D. M., J. Suppe, and F. A. Dahlen, Mechanics of fold-and-thrust belts and accretionary wedges, J. Geophys. Res., 88, 11531172, 1983.
  • Davis, E. E., et al., FlankFlux: An experiment to study the nature of hydrothermal circulation in young oceanic crust, Can. J. Earth Sci., 29, 925952, 1992.
  • Defant, M. J., and M. S. Drummond, Derivation of some modern arc magmas by melting of young subducted lithosphere, Nature, 347, 662665, 1990.
  • DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Current plate motions, Geophys. J. Int., 101, 425478, 1990.
  • DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein, Effect of recent revisions to the geomagnetic reversal time scale on estimates of current plate motions, Geophys. Res. Lett., 21, 21912194, 1994.
  • Dewey, J. F., and B. F. Windley, Growth and differentiation of the continental crust, Philos. Trans. R. Soc. London Ser. A, 301, 189206, 1981.
  • DeWit, M. J., and R. A. Hart, Earth's earliest continental lithosphere, hydrothermal flux and crustal recycling, Lithos, 30, 309335, 1993.
  • Dick, H. J. B., P. S. Meyer, S. H. Bloomer, S. H. Kirby, D. S. Stakes, and C. Mawer, Lithostratigraphic evolution of an in-situ section of oceanic layer 3, Proc. Ocean Drill. Program Sci. Results, 118, 439538, 1991.
  • Dick, H. J. B., et al., A long in situ section of the lower ocean crust: Results of ODP Leg 176 drilling at the Southwest Indian Ridge, Earth Planet. Sci. Lett., 179, 3151, 2000.
  • Dick, H. J., J. Lin, P. J. Michael, H. Schouten, and J. E. Snow, Ultra-slow spreading - a new class of ocean ridge, Eos Trans. AGU, 83(47), Fall Meet. Suppl., F1267, 2002.
  • Dickinson, W. R., Potash-depth (K-h) relations in continental margin and intra-oceanic magmatic arcs, Geology, 3, 5356, 1975.
  • Dickinson, W. R., and T. Hatherton, Andesitic volcanism and seismicity around the Pacific, Science, 157, 801803, 1967.
  • Donnelly, T. W., R. A. Pritchard, R. Emmermann, and H. Puchelt, The aging of oceanic crust: synthesis of the mineralogical and chemical results of Deep Sea Drilling Project, Legs 51 through 53, Initial Rep. Deep Sea Drill. Proj., 51–53, 15631577, 1979a.
  • Donnelly, T. W., G. Thompson, and P. T. Robinson, Very-low-temperature hydrothermal alteration of the oceanic crust and the problem of fluxes of potassium and magnesium, in Deep Drilling Results in the Atlantic Ocean: Ocean Crust, Maurice Ewing Ser., vol. 2, edited by M. Talwani, C. G. A. Harrison, and D. E. Hayes, pp. 369382, AGU, Washington, D.C., 1979b.
  • Dymond, J., and M. Lyle, Particle fluxes in the ocean and implications for sources and preservation of ocean sediments, in Material Fluxes on the Surface of the Earth, edited by W. W. Hay et al., pp. 125142, Natl. Acad. Press, Washington, D.C., 1994.
  • Ekart, D., T. E. Cerling, I. P. Montanez, and N. J. Tabor, A 400 million year carbon isotope record of pedogenic carbonate: Implications for paleoatmospheric carbon dioxide, Am. J. Sci., 299, 805827, 1999.
  • Elderfield, H., and A. Schultz, Mid-ocean ridge hydrothermal fluxes and the chemical composition of the ocean, Annu. Rev. Earth Planet. Sci., 24, 191224, 1996.
  • Ernewein, M., C. Pflumino, and H. Whitechurch, The death of an accretion zone as evidenced by the magmatic history of the Samail ophiolite (Oman), Tectonophysics, 151, 247274, 1988.
  • Ernst, W. G., Hornblende, the continent maker: Evolution of H2O during circum-Pacific subduction versus continental collision, Geology, 27, 675678, 1999.
  • Fehn, U., and L. M. Cathles, The influence of plate movement on the evolution of hydrothermal convection cells in the oceanic crust, Tectonophysics, 125, 289312, 1986.
  • Fisher, A. T., Permeability within basaltic oceanic crust, Rev. Geophys., 36, 143182, 1998.
  • Fisher, A. T., and K. Becker, Channelized fluid flow in oceanic crust reconciles heat-flow and permeability data, Nature, 403, 7174, 2000.
  • Fisher, A., K. Becker, T. N. Narasimhan, M. Langseth, and M. Mottl, Passive, off-axis convection on the southern flank of the Costa Rica Rift, J. Geophys. Res., 95, 93439370, 1990.
  • Fisher, A. T., K. Becker, and T. N. Narasimhan, Off-axis hydrothermal circulation: Parametric tests of a refined model of processes at Deep Sea Drilling Project/Ocean Drilling Program Site 504, J. Geophys. Res., 99, 30973121, 1994.
  • Fisher, D. M., Fabrics and veins in the forearc: A record of cyclic fluid flow at depths of <15 km, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 7589, AGU, Washington, D.C., 1996.
  • Fox, P. J., and J. Stroup, The plutonic foundation of the oceanic crust, in The Sea, vol. 7, The Oceanic Lithosphere, edited by C. Emiliani, pp. 119218, John Wiley, New York, 1981.
  • Francheteau, J., P. Choukroune, R. Hekinian, X. Le Pichon, and H. D. Needham, Oceanic fracture zones do not provide deep sections in the crust, Can. J. Earth Sci., 13, 12231235, 1975.
  • Franck, S., and C. Bounama, Continental growth and volatile exchange during Earth's evolution, Phys. Earth Planet. Int., 100, 189196, 1997.
  • François, L. M., and J. C. G. Walker, Modelling the Phanerozoic carbon cycle and climate: Constraints from the 87Sr/86Sr isotopic ratio of seawater, Am. J. Sci., 292, 81135, 1992.
  • Furnes, H., and H. Staudigel, Biological mediation in ocean crust alteration: How deep is the deep biosphere? Earth Planet. Sci. Lett., 166, 97103, 1999.
  • Furnes, H., H. Staudigel, I. H. Thorseth, T. Torsvik, K. Muehlenbachs, and O. Tumyr, Bioalteration of basaltic glass in the oceanic crust, Geochem. Geophys. Geosys., 2, Paper number 2000GC000150, 2001.
  • Fyfe, W. S., Deep fluids and volatile recycling; crust to mantle, Tectonophysics, 275, 243251, 1997.
  • Gaffin, S., Ridge volume dependence on seafloor generation rate and inversion using long term sealevel change, Am. J. Sci., 287, 596611, 1987.
  • Gerlach, T. M., Degassing of carbon dioxide from basaltic magma at spreading centers: 2. mid-oceanic ridge basalts, J. Volcanol. Geotherm. Res., 39, 221232, 1989.
  • Gill, J., Orogenic Andesites and Plate Tectonics, 390 pp., Springer-Verlag, New York, 1981.
  • Gillis, K., and P. T. Robinson, Distribution of alteration zones in the upper oceanic crust, Geology, 16, 262266, 1988.
  • Goldberg, D., The role of downhole measurements in marine geology and geophysics, Rev. Geophys., 35, 315342, 1997.
  • Green, N. L., and D. L. Harry, On the relationship between subducted slab age and arc basalt petrogenesis, Cascadia subduction system, North America, Earth Planet. Sci. Lett., 171, 367381, 1999.
  • Grevemeyer, I., and W. Weigel, Seismic velocities of the uppermost igneous crust versus age, Geophys. J. Int., 124, 631635, 1996.
  • Hacker, B. R., Eclogite formation and the rheology, buoyancy, seismicity, and H2O content of oceanic crust, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 337346, AGU, Washington, D.C., 1996.
  • Hamano, Y., Physical properties of basalts from Holes 417D and 418A, Initial Rep. Deep Sea Drill. Proj., 51–53, 14571466, 1979.
  • Harrison, C. G. A., Constraints on ocean volume change since the Archean, Geophys. Res. Lett., 26, 19131916, 1999.
  • Harry, D. L., and N. L. Green, Slab dehydration and basalt petrogenesis in subduction systems involving very young oceanic lithosphere, Chem. Geol., 160, 309333, 1999.
  • Hart, R. A., Chemical exchange between sea water and deep ocean basalts, Earth Planet. Sci. Lett., 9, 269279, 1970.
  • Hart, R. A., A model for chemical exchange in the basalt-seawater system of oceanic Layer II, Can. J. Earth Sci., 10, 799816, 1973.
  • Hart, S. R., K, Rb, Cs contents, and K/Rb, K/Cs ratios of fresh and altered submarine basalts, Earth Planet. Sci. Lett., 6, 295303, 1969.
  • Hart, S. R., and H. Staudigel, The control of alkalis and uranium in seawater by ocean crust alteration, Earth Planet. Sci. Lett., 58, 202212, 1982.
  • Hart, S. R., A. J. Erlank, and E. J. D. Kable, Sea floor basalt alteration: some chemical and Sr isotopic effects, Contrib. Mineral. Petrol., 44, 219230, 1974.
  • Hart, S. R., J. Blusztajn, H. J. B. Dick, P. S. Meyer, and K. Muehlenbachs, The fingerprint of seawater circulation in a 500-meter section of ocean crust gabbros, Geochim. Cosmochim. Acta, 63, 40594080, 1999.
  • Hay, W. W., Pleistocene-Holocene fluxes are not the Earth's norm, in Material Fluxes on the Surface of the Earth, edited by W. W. Hay et al., pp. 1527, Natl. Acad. Press, Washington, D.C., 1994.
  • Hays, J. D., and W. C. Pitman III, Lithospheric plate motion, sea-level changes, and climatic and ecological consequences, Nature, 246, 1822, 1973.
  • Hilde, T. W. C., Sediment subduction versus accretion around the Pacific, Tectonophysics, 99, 381397, 1983.
  • Hill, M. N., Recent geophysical exploration of the ocean floor, Phys. Chem. Earth, 2, 129163, 1957.
  • Holland, H. D., River transport to the oceans, in The Sea, vol. 7, The Oceanic Lithosphere, edited by C. Emiliani, pp. 763800, John Wiley, New York, 1981.
  • Honnorez, J., The aging of the oceanic crust at low temperature, in The Sea, vol. 7, The Oceanic Lithosphere edited by C. Emiliani, pp. 525587, John Wiley, New York, 1981.
  • Honnorez, J., C. Laverne, H.-W. Hubberten, R. Emmermann, and K. Muehlenbachs, Alteration processes in layer 2 basalts from Deep Sea Drilling Project Hole 504B, Costa Rica Rift, Initial Rep. Deep Sea Drill. Proj., 69, 509546, 1983.
  • Honza, E., H. L. Davies, J. B. Keene, and D. L. Tiffin, Plate boundaries and evolution of the Solomon Sea region, Geo. Mar. Lett., 7, 161168, 1987.
  • Humphris, S. E., Hydrothermal processes at mid-ocean ridges, U. S. Natl. Rep. Int. Union Geod. Geophys., 1991–1994, Rev. Geophys., 33, 7180, 1995.
  • Hyndman, R. D., Dipping seismic reflectors, electrically conductive zones and trapped water in the crust over a subducting plate, J. Geophys. Res., 93, 13,39113,405, 1988.
  • Hyndman, R. D., and M. J. Drury, The physical properties of oceanic basement rocks from deep drilling on the Mid-Atlantic Ridge, J. Geophys. Res., 81, 40424052, 1976.
  • Hyndman, R. D., M. Yamano, and D. A. Oleskevich, The seismogenic zone of subduction thrust faults, Island Arc, 6, 244260, 1997.
  • Ildefonse, B., and P. Pezard, Electrical properties of slow-spreading ridge gabbros from ODP Site 735, Southwest Indian Ridge, Tectonophysics, 330, 6992, 2001.
  • Ionov, D. A., W. L. Griffin, and S. Y. Oreilly, Volatile-bearing minerals and lithophile trace elements in the upper mantle, Chem. Geol., 141, 153184, 1997.
  • Ito, E. D., M. Harris, and A. T. Anderson Jr., Alteration of oceanic crust and geologic cycling of chlorine and water, Geochim. Cosmochim. Acta, 47, 16131624, 1983.
  • Itturrino, G. J., N. I. Christensen, K. Becker, L. O. Boldreel, P. K. H. Harvey, and P. Pezard, Physical properties and elastic constants of upper crustal rocks from core-log measurements in Hole 504B, Proc. Ocean Drill. Program Sci. Results, 137/140, 273292, 1995.
  • Iwamori, H., Transportation of H2O and melting in subduction zones, Earth Planet. Sci. Lett., 160, 6580, 1998.
  • Iwamori, H., Transportation of H2O and melting beneath the Japan arcs, Bull. Earthquake Res. Inst. Univ. Tokyo, 76, 377389, 2001.
  • Jambon, A., B. Deruelle, G. Dreibus, and F. Pineau, Chlorine and bromine abundance in MORB; The contrasting behaviour of the Mid-Atlantic Ridge and East Pacific Rise and implications for chlorine geodynamic cycle, Chem. Geol., 126, 101117, 1995.
  • Jarrard, R. D., Relations among subduction parameters, Rev. Geophys., 24, 217284, 1986.
  • Jarrard, R. D., R. L. Larson, A. T. Fisher, and L. J. Abrams, Geophysical aging of oceanic crust: Evidence from Hole 801C, Proc. Ocean Drill. Program Sci. Results, 144, 649663, 1995.
  • Jarrard, R. D., L. J. Abrams, R. Pockalny, R. L. Larson, and T. Hirono, Physical properties of upper oceanic crust: ODP Hole 801C and the waning of hydrothermal circulation, J. Geophys. Res, 108(D7), 2188, doi:10.1029/2001JB001727, 2003.
  • Jochum, K. P., A. W. Hofmann, E. Ito, H. M. Seufert, and W. M. White, K, U and Th in mid-ocean ridge basalt glasses and heat production, K/U and K/Rb in the mantle, Nature, 306, 431436, 1983.
  • Johnson, D. M., Fluid permeability of oceanic basalts, Initial Rep. Deep Sea Drill. Proj., 51–53, 14731477, 1979a.
  • Johnson, D. M., Crack distribution in the upper oceanic crust and its effects upon seismic velocity, seismic structure, formation permeability, and fluid circulation, Initial Rep. Deep Sea Drill. Proj., 51–53, 14791490, 1979b.
  • Johnson, H. P., and S. W. Semyan, Age variation in the physical properties of oceanic basalts: implications for crustal formation and evolution, J. Geophys. Res., 99, 31233134, 1994.
  • Karato, S., Physical properties of basalts from Deep Sea Drilling Project Hole 504B, Initial Rep. Deep Sea Drill. Proj., 69, 687695, 1983a.
  • Karato, S., Physical properties of basalts from the Galapagos, Leg 70, Initial Rep. Deep Sea Drill. Proj., 70, 423428, 1983b.
  • Karson, J. A., Internal structure of oceanic lithosphere: a perspective from tectonic windows, in Faulting and Magmatism at Mid-Ocean Ridges, Geophys. Monogr. Ser., vol. 106, edited by W. R. Buck et al., 177218, AGU, Washington, D.C., 1998.
  • Karson, J. A., Geologic structure of the uppermost oceanic crust created at fast- to intermediate-rate spreading centers, Annu. Rev. Earth Planet. Sci., 30, 347384, 2002.
  • Kasahara, J., A. Kamimura, G. Fujie, and R. Hino, Influence of water on earthquake generation along subduction zones, Bull. Earthquake. Res. Inst. Univ. Tokyo, 76, 291303, 2001.
  • Kastner, M., H. Elderfield, and J. B. Martin, Fluids in convergent margins: What do we know about their composition, origin, role in diagenesis and importance for oceanic chemical fluxes, Philos. Trans. R. Soc. London, Ser. A, 335, 243259, 1991.
  • Kastner, M., et al., Geochemical and isotopic evidence for fluid flow in the western Nankai subduction zone, Japan, Proc. Ocean Drill. Program Sci. Results, 131, 397413, 1993.
  • Kato, T., and Y. Kotake, Tectonics of the Philippine Sea plate as seen from GPS observations, Eos Trans. AGU, 83(47), Fall Meet. Suppl., F1321, 2002.
  • Kerrick, D. M., Present and past nonanthropogenic CO2 degassing from the solid Earth, Rev. Geophys., 39, 565585, 2001.
  • Kerrick, D. M., and J. A. D. Connolly, Subduction of ophicarbonates and recycling of CO2 and H2O, Geology, 26, 375378, 1998.
  • Kerrick, D. M., and J. A. D. Connolly, Metamorphic devolatilization of subducted oceanic metabasalts: Implications for seismicity, arc magmatism and volatile recycling, Earth Planet. Sci. Lett., 189, 1929, 2001.
  • Kirby, S. H., E. R. Engdahl, and R. Denlinger, Intraslab earthquakes and arc volcanism: dual physical expressions of crustal and uppermost mantle metamorphism in subducting slabs, in Subduction Top to Bottom, Geophys. Monogr. Ser., 96, edited by G. E. Bebout et al., pp. 195214, AGU, Washington, D.C., 1996.
  • Kominz, M. A., Oceanic ridge volumes and sea-level change: An error analysis, in Interregional Unconformities and Hydrocarbon Accumulation, edited by J. S. Schlee, Mem. Am. Assoc. Petrol. Geol., 36, 109127, 1984.
  • Larson, R. L., Geological consequences of superplumes, Geology, 19, 547550, 1991.
  • Lawrence, J. R., J. J. Drever, and M. Kastner, Low temperature alteration of basalts predominates at DSDP Site 395, Initial Rep. Deep Sea Drill. Proj., 45, 609612, 1978.
  • Lecuyer, C., M. Brouxel, and F. Albarede, Elemental fluxes during hydrothermal alteration of the Trinity Ophiolite (California, USA) by seawater, Chem. Geol., 89, 87115, 1990.
  • Leeman, W. P., Boron and other fluid-mobile elements in volcanic arc lavas: Implications for subduction processes, in Subduction Top to Bottom, Geophys. Monogr., vol. 96, edited by G. E. Bebout et al., pp. 269276, AGU, Washington, D.C., 1996.
  • Le Pichon, X., P. Henry, and the Kaiko-Nankai Scientific Crew, Water budgets in accretionary wedges: A comparison, Philos. Trans. R. Soc. London Ser. A, 335, 315330, 1991.
  • Le Pichon, X., P. Henry, and S. Lallemant, Accretion and erosion in subduction zones: The role of fluids, Annu. Rev. Earth Planet. Sci., 21, 307331, 1993.
  • Lippard, S. J., A. W. Shelton, and I. G. Gass, The Ophiolite of Northern Oman, 178 pp., Blackwell Sci., Malden, Mass., 1986.
  • Liu, J., S. R. Bohlen, and W. G. Ernst, Stability of hydrous phases in subducting oceanic crust, Earth Planet. Sci. Lett., 143, 161171, 1996.
  • Livingstone, D. A., Chemical composition of rivers and lakes, U. S. Geol. Survey Prof. Paper, 440-G, G1G64, 1963.
  • Magenheim, A. J., A. J. Spivack, P. J. Michael, and J. M. Gieskes, Chlorine stable isotope composition of the oceanic crust: Implication for Earth's distribution of chlorine, Earth Planet. Sci. Lett., 131, 427432, 1995.
  • Martin, W. R., and F. L. Sayles, Seafloor diagenetic fluxes, in Material Fluxes on the Surface of the Earth, edited by W. W. Hay, et al., pp. 143163, Natl. Acad. Press, Washington, D.C., 1994.
  • Marty, B., and I. N. Tolstikhin, CO2 fluxes from midocean ridges, arcs and plumes, Chem. Geol., 145, 233248, 1998.
  • McCaffrey, R., Influences of recurrence times and fault zone temperatures on the age-rate dependence of subduction zone seismicity, J. Geophys. Res., 102, 22,83922,854, 1997.
  • McCulloch, M. T., The role of subducted slabs in an evolving earth, Earth Planet. Sci. Lett., 115, 89100, 1993.
  • McGovern, P. J., and G. Schubert, Thermal evolution of the Earth: Effects of volatile exchange between atmosphere and interior, Earth Planet. Sci. Lett., 96, 2737, 1989.
  • Meade, C., and R. Jeanloz, Deep-focus earthquakes and recycling of water into the Earth's mantle, Science, 252, 6872, 1991.
  • Meybeck, M., Concentrations des eaux fluviales en éléments majeurs et apports en solution aux océans, Rev. Geol. Dyn. Geogr. Phys., 21, 215246, 1979.
  • Meybeck, M., Atmospheric inputs and river transport of dissolved substances, in Dissolved Load of Rivers and Surface Water Quantity/Quality Relationships, Proceedings of Hamburg Symposium, IAHS Publ., 141, edited by B. W. Webb, pp. 173192, Int. Assoc. of Hydrol. Sci. Press, Wallingford, Oxfordshire, U.K., 1983.
  • Meybeck, M., Origin and variable composition of present day riverborne material, in Material Fluxes on the Surface of the Earth, edited by W. W. Hay et al., pp. 6173, Natl. Acad. Press, Washington, D.C., 1994.
  • Milliman, J. D., and R. H. Meade, World-wide delivery of river sediment to the oceans, J. Geology, 91, 121, 1983.
  • Milliman, J. D., and J. P. M. Syvitski, Geomorphic/tectonic control of sediment discharge to the ocean: the importance of small mountainous rivers, in Material Fluxes on the Surface of the Earth, edited by W. W. Hay, et al., pp. 7485, Natl. Acad. Press, Washington, D.C., 1994.
  • Minster, J. B., and T. H. Jordan, Present-day plate motions, J. Geophys. Res., 83, 53315354, 1978.
  • Miyazaki, S., and K. Heki, Crustal velocity field of southwest Japan: Subduction and arc-arc collision, J. Geophys. Res., 106, 43054326, 2001.
  • Moore, J. C., and D. Saffer, Updip limit of the seismogenic zone beneath the accretionary prism of southwest Japan: An effect of diagenetic to low-grade metamorphic processes and increasing effective stress, Geology, 29, 183186, 2001.
  • Moore, J. C., and P. Vrolijk, Fluids in accretionary prisms, Rev. Geophys., 30, 113135, 1992.
  • Moos, D., and D. Marion, Morphology of extrusive basalts and its relationship to seismic velocities in the shallow oceanic crust, J. Geophys. Res., 99, 29852994, 1994.
  • Mottl, M. J., and C. G. Wheat, Hydrothermal circulation through mid-ocean ridge flanks: Fluxes of heat and magnesium, Geochim. Cosmochim. Acta, 58, 22252237, 1994.
  • Muehlenbachs, K., The alteration and aging of the basaltic layer of the sea floor: oxygen isotope evidence from DSDP/IPOD Legs 51, 52, and 53, Initial Rep. Deep Sea Drill. Proj., 51–53, 11591167, 1979.
  • Nakamori, T., Global carbonate accumulation rates from Cretaceous to Present and their implications for the carbon cycle model, Island Arc, 10, 18, 2001.
  • Nockolds, S. R., Average chemical compositions of some igneous rocks, Geol. Soc. Am. Bull., 65, 10071032, 1954.
  • Nur, A., Effects of stress on velocity anisotropy in rocks with cracks, J. Geophys. Res., 76, 20222034, 1971.
  • Omori, S., S. Kamiya, S. Maruyama, and D. Zhao, Morphology of the intraslab seismic zone and devolatilization phase equilibria of the subducting slab peridotite, Bull. Earthquake Res. Inst. Univ. Tokyo, 76, 455478, 2001.
  • Ono, S., Stability limits of hydrous minerals in sediment and midocean ridge basalt compositions: Implications for water transport in subduction zones, J. Geophys. Res., 103, 18,25318,267, 1998.
  • Parsons, B., The rates of plate creation and consumption, Geophys. J. R. Astron. Soc., 67, 437448, 1981.
  • Pawley, A. R., The pressure and temperature stability limits of lawsonite: Implications for H2O recycling in subduction zones, Contrib. Mineral. Petrol., 118, 99108, 1994.
  • Pawley, A. R., and J. R. Holloway, Water sources for subduction zone volcanism: New experimental constraints, Science, 260, 664667, 1993.
  • Peacock, S. M., Fluid processes in subduction zones, Science, 248, 329337, 1990.
  • Peacock, S. M., Numerical simulation of subduction zone pressure temperature time paths - constraints on fluid production and arc magmatism, Philos. Trans. R. Soc. London Ser. A, 335, 341353, 1991.
  • Peacock, S. M., The importance of blueschist-eclogite dehydration reactions in subducting oceanic crust, Geol. Soc. Amer. Bull., 105, 684694, 1993.
  • Peacock, S. M., Thermal and petrological structure of subduction zones, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 119133, AGU, Washington, D.C., 1996.
  • Peacock, S. M., Are the lower planes of double seismic zones caused by serpentine dehydration in subducting oceanic mantle? Geology, 29, 299302, 2001.
  • Peacock, S. M., and R. D. Hyndman, On the importance of serpentine in the subduction factory, Eos Trans. AGU, 82(47), Fall Meet. Suppl., F1154, 2001.
  • Peacock, S. M., T. Rushmer, and A. B. Thompson, Partial melting of subducting oceanic crust, Earth Planet. Sci. Lett., 121, 227244, 1994.
  • Pearson, P. N., and M. R. Palmer, Atmospheric carbon dioxide concentrations over the past 60 million years, Nature, 406, 695699, 2000.
  • Pezard, P. A., Electrical properties of mid-ocean ridge basalt and implications for the structure of the upper oceanic crust at Hole 504B, J. Geophys. Res., 95, 92379264, 1990.
  • Pezard, P. A., F. Einaudi, D. Hermitte, J.-J. Cocheme, C. Coulon, and C. Laverne, MORB emplacement and structure; Insights from the Semail Ophiolite, Oman, Geophys. Res. Lett., 27, 39333936, 2000.
  • Philippot, P., P. Agrinier, and M. Scambelluri, Chlorine cycling during subduction of altered oceanic crust, Earth Planet. Sci. Lett., 161, 3344, 1998.
  • Plank, T., and C. H. Langmuir, Tracing trace elements from sediment input to volcanic output at subduction zones, Nature, 362, 739743, 1993.
  • Plank, T., and C. H. Langmuir, The chemical composition of subducting sediment and its consequences for the crust and mantle, Chem. Geol., 145, 325394, 1998.
  • Poli, S., and M. W. Schmidt, H2O transport and release in subduction zones: Experimental constraints on basaltic and andesitic systems, J. Geophys. Res., 100, 22,29922,314, 1995.
  • Purdy, G. M., L. S. L. Kong, G. L. Christeson, and S. C. Solomon, Relationship between spreading rate and the seismic structure of mid-ocean ridges, Nature, 355, 815817, 1992.
  • Raitt, R. W., The crustal rocks, in The Sea, vol. 3, edited by M. N. Hill, pp. 85102, Wiley-Interscience, New York, 1963.
  • Raleigh, C. B., and M. S. Paterson, Experimental deformation of serpentinite and its tectonic implications, J. Geophys. Res., 70, 39653985, 1965.
  • Rangin, C., and E. Silver, Geological setting of the Celebes and Sulu Seas, Proc. Ocean Drill Program Initial Rep., 124, 3542, 1990.
  • Ransom, B., and H. C. Helgeson, A chemical and thermodynamic model of dioctehedral 2:1 layer clay minerals in diagenetic processes: Dehydration of dioctahedral aluminous smectite as a function of temperature and depth in sedimentary basins, Am. J. Sci., 295, 245281, 1995.
  • Rea, D. K., Geologic records in deep-sea muds, GSA Today, 3, 205210, 1993.
  • Rea, D. K., and L. J. Ruff, Composition and mass flux of sediment entering the world's subduction zones: Implications for global sediment budgets, great earthquakes, and volcanism, Earth Planet. Sci. Lett., 140, 112, 1996.
  • Reymer, A., and G. Schubert, Phanerozoic addition rates to the continental crust and crustal growth, Tectonics, 3, 6377, 1984.
  • Robertson, A. H. F., and Shipboard Scientific Party, Tectonic introduction, Proc. Ocean Drill. Program Init. Rep., 160, 520, 1996.
  • Robinson, P. T., M. F. Flower, H. Staudigel, and D. A. Swanson, Lithology and eruptive stratigraphy of Cretaceous oceanic crust, western Atlantic, Initial Rep. Deep Sea Drill. Proj., 51–53, 15351556, 1979.
  • Robinson, P. T., H. J. B. Dick, D. Von Herzen, and R. Pierre, Metamorphism and alteration in oceanic layer 3: Hole 735B, Proc. Ocean Drill. Program Sci. Results, 118, 541562, 1991.
  • Ronov, A. B., and A. A. Yaroshevskiy, A new model for the chemical structure of the Earth's crust, Geochem. Int., 13, 89121, 1976.
  • Rubey, W. W., Geologic history of sea water: an attempt to state the problem, Geol. Soc. Am. Bull., 62, 11111147, 1951.
  • Ruff, L., and H. Kanamori, Seismicity and the subduction process, Phys. Earth Planet. Int., 23, 240252, 1980.
  • Ruff, L. J., and B. W. Tichelaar, What controls the seismogenic plate interface in subduction zones? in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 105111, AGU, Washington, D.C., 1996.
  • Ryan, J., J. Morris, G. Bebout, and B. Leeman, Describing chemical fluxes in subduction zones: insights from “depth profiling” studies of arc and forearc rocks, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 263268, AGU, Washington, D.C., 1996.
  • Saffer, D. M., and B. A. Bekins, Episodic fluid flow in the Nankai accretionary complex: Timescale, geochemistry, flow rates, and fluid budget, J. Geophys. Res., 103, 30,35130,370, 1998.
  • Saffer, D. M., and B. A. Bekins, Hydrologic controls on the morphology and mechanics of accretionary wedges, Geology, 30, 271274, 2002.
  • Sager, W., Argo abyssal plain magnetic lineations revisited: Implications for the onset of seafloor spreading and tectonic evolution of the Eastern Indian Ocean, Proc. Ocean Drill. Program Sci. Results, 123, 659669, 1992.
  • Salisbury, M. H., N. I. Christensen, and R. H. Wilkens, Nature of the Layer 2/3 transition from a comparison of laboratory and logging velocities and petrology at the base of Hole 504B, Proc. Ocean Drill. Program Sci. Results, 148, 409414, 1996.
  • Sansone, F. J., M. J. Mottl, E. J. Olson, C. G. Wheat, and M. D. Lilley, CO2-depleted fluids from mid-ocean ridge-flank hydrothermal springs, Geochim. Cosmochim. Acta, 62, 22472252, 1998.
  • Schilling, J.-G., C. K. Unni, and M. L. Bender, Origin of chlorine and bromine in the oceans, Nature, 273, 631636, 1978.
  • Schmidt, M. W., Experimental constraints on recycling of potassium from subducted oceanic crust, Science, 272, 19271930, 1996.
  • Schmidt, M. W., and S. Poli, Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation, Earth Planet. Sci. Lett., 163, 361379, 1998.
  • Schmincke, H.-U., and U. Bednarz, Pillow-, sheet flow-, and breccia volcano-tectonic-hydrothermal cycles in the Extrusive Series of the northwestern Troodos Ophiolite (Cyprus), in Proceedings Symposium Troodos 87–Ophiolites and Oceanic Lithosphere, edited by J. Malpas et al., pp. 185207, Geol. Surv. Dept., Nicosia, Cyprus, 1990.
  • Schreiber, E., and P. J. Fox, Compressional wave velocities and mineralogy of fresh basalts from the FAMOUS area and the Oceanographer Fracture Zone and the texture of Layer 2A of the oceanic crust, J. Geophys. Res., 81, 40714076, 1976.
  • Schreiber, E., and P. J. Fox, Density and P-wave velocity of rocks from the FAMOUS region and their implication to the structure of the oceanic crust, Geol. Soc. Am. Bull., 88, 600608, 1977.
  • Seno, T., S. Stein, and A. E. Gripp, A model for the motion of the Philippine Sea plate consistent with NUVEL-1 and geological data, J. Geophys. Res., 98, 17,94117,948, 1993.
  • Serra, O., Fundamentals of Well Log Interpretation, vol. 2, The Interpretation of Logging Data, 684 pp., Elsevier, New York, 1986.
  • Shaw, P., Age variations of oceanic crust Poisson's ratio: Inversion and a porosity inversion model, J. Geophys. Res., 99, 30573066, 1994.
  • Shimakawa, Y., and Y. Honkura, Electrical conductivity structure beneath the Ryukyu Trench-Arc system and its relation to subduction of the Philippine Sea plate, J. Geomag. Geoelect., 43, 120, 1991.
  • Shipboard Scientific Party, Site 395, Initial Rep. Deep Sea Drill. Proj., 45, 131264, 1978.
  • Shipboard Scientific Party, Site 562, Initial Rep. Deep Sea Drill. Proj., 82, 255278, 1985.
  • Shipboard Scientific Party, Site 735, Proc. Ocean Drill. Program Initial Rep., 118, 89222, 1989.
  • Shipboard Scientific Party, Site 768, Proc. Ocean Drill. Program Initial Rep., 124, 195297, 1990a.
  • Shipboard Scientific Party, Site 770, Proc. Ocean Drill. Program Initial Rep., 124, 343397, 1990b.
  • Shipboard Scientific Party, Site 504, Proc. Ocean Drill. Program Initial Rep., 140, 37200, 1992.
  • Shipboard Scientific Party, Site 894, Proc. Ocean Drill. Program Initial Rep., 147, 45108, 1993.
  • Shipboard Scientific Party, Site 735 [online], Proc. Ocean Drill. Program Initial Rep., 176, 1314, 1999. (Available at
  • Shipboard Scientific Party, Site 801 [online], Proc. Ocean Drill. Program Initial Rep., 185, 1222, 2000a. (Available at
  • Shipboard Scientific Party, Site 1149 [online], Proc. Ocean Drill. Program Initial Rep., 185, 1190, 2000b. (Available at
  • Shipboard Scientific Party, Site 1222 [online], Proc. Ocean Drill. Program Initial Rep., 199, 157, 2002. (Available at
  • Sleep, N. H., Thermal history and degassing of the Earth; some simple calculations, J. Geol., 87, 671686, 1979.
  • Sleep, N. H., and T. J. Wolery, Sediment subduction and alteration of oceanic crust, a chemical constraint (abstract), Eos Trans. AGU, 60, 392, 1979.
  • Small, C., Global systematics of mid-ocean ridge morphology, in Faulting and Magmatism at Mid-Ocean Ridges, Geophys. Monogr. Ser., vol. 106, edited by W. R. Buck et al., pp. 125, AGU, Washington, D.C., 1998.
  • Sobolev, A. V., and M. Chaussidon, H2O concentrations in primary melts from supra-subduction zones and midocean ridges: Implications for H2O storage and recycling in the mantle, Earth Planet. Sci. Lett., 137, 4555, 1996.
  • Spivack, A. J., and H. Staudigel, Low-temperature alteration of the upper oceanic crust and the alkalinity budget of seawater, Chem. Geol., 115, 239247, 1994.
  • Sprague, D., and H. N. Pollack, Heat flow in the Mesozoic and Cenozoic, Nature, 285, 393395, 1980.
  • Staudigel, H., and S. D. King, Ultrafast subduction: The key to slab recycling efficiency and mantle differentiation, Earth Planet. Sci. Lett., 109, 517530, 1992.
  • Staudigel, H., S. R. Hart, and S. Richardson, Alteration of the oceanic crust: Processes and timing, Earth Planet. Sci. Lett., 52, 311327, 1981.
  • Staudigel, H., S. R. Hart, H.-U. Schmincke, and B. M. Smith, Cretaceous ocean crust at DSDP sites 417 and 418; Carbon uptake from weathering versus loss by magmatic outgassing, Geochim. Cosmochim. Acta, 53, 30913094, 1989.
  • Staudigel, H., T. Plank, B. White, and H.-U. Schmincke, Geochemical fluxes during seafloor alteration of the basaltic upper oceanic crust; DSDP sites 417 and 418, in Subduction Top to Bottom, Geophys. Monogr. Ser., vol. 96, edited by G. E. Bebout et al., pp. 1938, AGU, Washington, D.C., 1996.
  • Staudigel, H., et al., Geochemical Earth Reference Model (GERM); Description of the initiative, Chem. Geol., 145, 153159, 1998a.
  • Staudigel, H., A. Yayanos, R. Chastain, G. Davies, E. A. T. Verdurmen, P. Schiffman, R. Bourcier, and H. de Baar, Biologically mediated dissolution of volcanic glass in seawater, Earth Planet. Sci. Lett., 164, 233244, 1998b.
  • Stein, C. A., and S. Stein, Constraints on hydrothermal heat flux through the oceanic lithosphere from global heat flow, J. Geophys. Res., 99, 30813095, 1994.
  • Stephen, R. A., Data report: physical properties measurements in ODP Hole 735B [online], Proc. Ocean Drill Program Sci. Results, 176, 2001. (Available at
  • Sugisaki, R., Chemical characteristics of volcanic rocks: Relation to plate movements, Lithos, 9, 1730, 1976.
  • Tatsumi, Y., and T. Kogiso, Trace-element transport during dehydration processes in the subducted oceanic crust: 2. Origin of chemical and physical characteristics in arc magmatism, Earth Planet. Sci. Lett., 148, 207221, 1997.
  • Thompson, A. B., Water in the Earth's upper mantle, Nature, 358, 295302, 1992.
  • Thompson, G., Metamorphic and hydrothermal processes: basalt-seawater interaction, in Oceanic Basalts, edited by P. A. Floyd, pp. 148173, Van Nostrand Reinhold, New York, 1991.
  • Toh, H., Electrical conductivity structure of the Izu-Bonin Arc revealed by seafloor electromagnetic observations, Ph.D. thesis, Ocean Research Inst., Univ. of Tokyo, Tokyo, 1993.
  • Tregoning, P., F. Tan, J. Gilliland, H. McQueen, and K. Lambeck, Present-day crustal motion in the Solomon Islands from GPS observations, Geophys. Res. Lett., 25, 36273630, 1998.
  • Ulmer, P., and V. Trommsdorff, Serpentine stability to mantle depths and subduction-related magmatism, Science, 268, 858861, 1995.
  • Underwood, M. B., et al., Sediment geochemistry, clay mineralogy, and diagenesis: A synthesis of data from Leg 131, Nankai Trough, Proc. Ocean Drill. Program Sci. Results, 131, 343363, 1993.
  • Vera, E. E., and J. C. Mutter, Crustal structure in the ROSE area of the East Pacific Rise: One-dimensional travel time inversion of sonobuoys and expanded spread profiles, J. Geophys. Res., 93, 66356648, 1988.
  • Von Herzen, R., Geothermal evidence for continuing hydrothermal circulation in older (>60 M.y.) ocean crust, in Hydrology of the Oceanic Lithosphere, edited by E. Davis, and H. Elderfield, Cambridge Univ. Press, New York, in press, 2003.
  • von Huene, R., and D. W. Scholl, Observations at convergent margins concerning sediment subduction, subduction erosion, and the growth of continental crust, Rev. Geophys., 29, 279316, 1991.
  • Wang, K., and E. E. Davis, In the pursuit of hydrothermal circulation and hydration depths of oceanic plates, Eos Trans. AGU, 83(47), Fall Meet. Suppl., F1252, 2002.
  • Wannamaker, P. E., J. R. Booker, A. G. Jones, A. D. Chave, J. H. Filloux, H. S. Waff, and L. K. Law, Resistivity cross section through the Juan de Fuca subduction system and its tectonic implications, J. Geophys. Res., 94, 14,12714,144, 1989.
  • White, R. S., D. McKenzie, and R. K. O'Nions, Oceanic crustal thickness from seismic measurements and rare earth element inversions, J. Geophys. Res., 97, 19,68319,715, 1992.
  • Wilkens, R. H., N. I. Christensen, and L. Slater, High-pressure seismic studies of Leg 69 and 70 basalts, Initial Rep. Deep Sea Drill. Proj., 69, 683686, 1983.
  • Wilkens, R. H., G. J. Fryer, and J. Karsten, Evolution of porosity and seismic structure of upper oceanic crust: importance of aspect ratios, J. Geophys. Res., 96, 17,98117,995, 1991.
  • Williams, Q., and R. J. Hemley, Hydrogen in the deep earth, Ann. Rev. Earth Planet. Sci., 29, 365418, 2001.
  • Wyllie, M. R. J., A. R. Gregory, and G. H. F. Gardner, Elastic wave velocities in heterogeneous and porous media, Geophysics, 21, 4170, 1956.
  • Zhang, Y., and A. Zindler, Distribution and evolution of carbon and nitrogen in Earth, Earth Planet. Sci. Lett., 117, 331345, 1993.
  • Zhou, W., D. R. Peacor, J. C. Alt, R. Van der Voo, and L.-S. Kao, TEM study of the alteration of interstitital glass in MORB by inorganic processes, Chem. Geol., 174, 365376, 2001a.
  • Zhou, W., R. Van der Voo, D. R. Peacor, D. Wang, and Y. Zhang, Low-temperature oxidation in MORB of titanomagnetite to titanomaghemite: A gradual process with implications for marine magnetic anomaly amplitudes, J. Geophys. Res., 106, 64096421, 2001b.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Age Dependence of Upper Crustal Properties
  5. 3. Data Set: Global Subduction
  6. 4. Subduction Fluxes
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
ggge300-sup-0001-t01.txtQuickTime video2KTab-delimited Table 1.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.