Previous   |    Next

doi:10.2204/iodp.proc.335.104.2012

The case for deep drilling of intact ocean crust

Drilling a complete in situ section of ocean crust has been an unfulfilled ambition of Earth scientists for many decades and provided the impetus for the conception of scientific deep ocean drilling. The production of new crust at mid-ocean ridges lays the foundation of the plate tectonic cycle and is a dominant process that has resurfaced >60% of our present-day planet since the Early Jurassic (<200 Ma). Magma eruption and intrusion, along with ocean floor hydrothermal exchange, are the principal mechanisms of heat and material transfer from the mantle to the crust, oceans, and atmosphere. The ocean crust is an environment of steep thermal, physical, and chemical gradients potentially with many of the ingredients required to initiate primordial life, as there is growing evidence for an enduring, active subsurface basalt-hosted microbial biosphere (e.g., Fisk et al., 1998; Bach and Edwards, 2003; Santelli et al., 2008; Rouxel et al., 2008; McLoughlin et al., 2009; McCarthy et al., 2011). Evidence for microbial activity was also recently reported in ~1 m.y. old gabbros collected during Integrated Ocean Drilling Program (IODP) Expedition 304/305 (Mason et al., 2010). Chemical exchanges between the ocean and crust over a wide range of temperatures exert major controls on seawater chemistry and partially buffer inputs from the erosion and weathering of continents brought to the oceans by rivers, glaciers, and groundwater (e.g., Palmer and Edmond, 1989; Vance et al., 2009).

Unfortunately, many of the key questions regarding the formation and evolution of the oceanic crust that are primary scientific goals of the IODP Initial Science Plan and numerous forerunner questions remain unanswered despite 50 years of scientific ocean drilling. This is principally due to the cursory sampling of the ocean crust, and in particular an absence of continuous deep crustal sections (see Wilson, Teagle, Acton, et al., 2003; Teagle et al., 2004; Dick et al., 2006; Ildefonse et al., 2007c). These fundamental questions remain compelling and increasingly relevant to understanding the wider Earth system with the growing appreciation of the interdependency between geological, climatic, and biogeochemical cycles.

Why study crust forming at fast spreading rates?

The vast majority (~70%) of magma derived from the mantle is brought into the Earth’s crust at the mid-ocean ridges, and approximately two-thirds of that magma cools and crystallizes in the lower portion of the oceanic crust. Seismic, bathymetric, and marine geological observations indicate that ocean crust formed at fast spreading rates (full rate > 80 mm/y) is much less variable than crust formed at slow spreading rates (<40 mm/y) and is closer to the ideal “Penrose” pseudostratigraphy developed from ophiolites (Anonymous, 1972). Hence, extrapolating fast-spreading accretion processes from a few sites might reasonably describe a significant portion of the Earth’s surface. Although <20% of modern ridges are moving apart at fast spreading rates (Fig. F1), nearly 50% of present-day ocean crust and ~30% of the Earth’s surface was produced at this pace of spreading (Fig. F2). The great majority of crust subducting into the mantle over the past ~200 m.y. formed at fast-spreading ridges (Müller et al., 2008), making characterizing this style of crust most relevant for understanding the recycling of crustal and ocean-derived components back into the mantle.

The spreading rate of the oceanic lithosphere has profound effects on the style of crustal accretion at mid-ocean ridges because of changing balances between plate motion, magma production, conductive and hydrothermal cooling, detachment tectonics, and serpentinization of the upper mantle (e.g., Dick, 1989; Cannat et al., 1995, 2004, 2009; Chen and Phipps Morgan, 1996; Dick et al., 2003; Escartin et al., 2008). Although insights on the formation of intrusive crust at detachment-dominated, slow-spread lithosphere have been obtained (Ocean Drilling Program [ODP] Legs 118, 153, 176, and 209 and IODP Site U1309; e.g., Dick et al., 2000; Blackman, Ildefonse, John, Ohara, Miller, MacLeod, and the Expedition 304/305 Scientists, 2006; Kelemen, Kikawa, Miller, et al., 2004; Ildefonse et al., 2007a; Blackman et al., 2011), the thermal regime and the melt supply and delivery in these settings differs significantly from those of the axial zone in fast-spreading lithosphere. Detailed understanding of the relatively uniform mechanisms operating at fast-spreading ridges would provide a vital benchmark against which heterogeneous accretion at slow-spreading ridges could be compared.

The need for basic geologic observations of ocean crustal architecture

Basic observations regarding the architecture of in situ present-day ocean crust, including rock types, geochemistry, and thicknesses of the volcanic, dike, and plutonic sections, are yet to be made. It is a fundamental weakness of our knowledge of the ocean crust that we are as yet unable to relate seismic and magnetic imaging of the ocean crust and geochemical inferences to basic geologic observations. We do not have a predictive understanding of the factors controlling the thicknesses of seismic and geological layers in the oceanic crust, which greatly precludes our ability to interpret regional geophysical data in geological terms. Drilling a few deep drill holes into intact ocean crust and studying samples having a range of seismic behaviors could greatly increase the confidence with which we interpret geophysical data and its use as a three-dimensional regional mapping tool (e.g., Fig. F3). Earth scientists often loosely speak of “Layer 3” when referring to the plutonic rocks of the ocean crust. However, the geological meaning and physical causes of the transition from seismic Layer 2 to Layer 3 velocities remain poorly understood. In Deep Sea Drilling Project (DSDP) Hole 504B, the only place where the Layer 2–3 transition has been penetrated in situ, the Layer 2–3 transition occurs near the middle of the ~1 km thick sheeted dike complex, where the transition to gabbroic rocks is at least 600 m deeper in the crust (Alt, Kinoshita, Stokking, et al., 1993; Detrick et al., 1994). At Site 504 the change from Layer 2 to Layer 3 appears to be related to changes in the secondary hydrothermal mineralogy (Alt et al., 1996) and/or crack porosity (Carlson, 2010). Whether this observation from the intermediate spreading rate crust sampled in Hole 504B is applicable to other spreading rates or ocean crust in general is yet to be tested.

Marine magnetic anomalies are one of the key observations that led to the development of plate tectonic theory, through the recognition that the ocean crust records the changing polarity of the Earth’s magnetic field through time (Vine and Matthews, 1963). Micrometer-sized grains of titanomagnetite within the erupted basalt are generally accepted to be the principal recorders of marine magnetic anomalies, but recent studies of tectonically exhumed lower crustal rocks and serpentinized upper mantle indicate that these deeper rocks may also be a significant source of the magnetic anomaly signal (Kikawa and Ozawa, 1992; Pariso and Johnson, 1993; Shipboard Scientific Party, 1999; Gee and Kent, 2007). Whether these deeper rocks have a significant influence on the magnetic field in undisrupted crust is unknown, as is the extent of secondary magnetite growth in gabbros and mantle assemblages away from transform faults. Sampling the plutonic layers of the crust could refine the Vine-Matthews hypothesis by characterizing the magnetic properties of gabbros and peridotites through drilling intact ocean crust, on a well-defined magnetic stripe, away from transform faults.

The most prominent melt feature observed by multichannel seismic experiments at fast-spreading mid-ocean ridges is a low-velocity zone some tens of meters thick, hundreds of meters across axis, and commonly continuous for many hundreds of kilometers along axis (e.g., Kent et al., 1994). This low-velocity zone is interpreted to be a dominantly magma rich lens (e.g., Detrick et al., 1987; Vera et al., 1990; Hussenoeder et al., 1996; Singh et al., 1998) that overlies a lower crustal region of reduced P- and S-wave velocities interpreted to be a hot crystal mush zone containing no more than a few percent of interstitial melt (e.g., Caress et al., 1992; Sinton and Detrick, 1992; Dunn et al., 2000). The roles of the low-velocity zone and axial magma lens in constructing fast-spreading ocean crust remain controversial. A family of elegant thermally based numerical models attempts to build the lower crust from the continuous subsidence of cumulate layers formed at the base of the axial melt lens (Fig. F4) (Sleep, 1975; Henstock et al., 1993; Phipps-Morgan and Chen, 1993; Quick and Denlinger, 1993). These models have major implications for the composition and deformation of the lower crust, but many of these predictions are not borne out by observations in ophiolites or the limited fast-spread plutonic ocean crust drilled to date. For example, petrologic observations from Hess Deep suggest that the uppermost gabbros, interpreted to represent the axial melt lens that formed the crust, are late-stage melt fractions, even more differentiated than erupted mid-ocean-ridge basalt (MORB), and question the significance of the axial melt lens in the formation of the lower oceanic crust (e.g., Natland and Dick, 2009).

The itinerary of melt formed by the partial melting of the mantle to its eruption on the seafloor remains poorly understood. For more than two decades it has been assumed that the compositions of MORB erupted onto the ocean floor can be interpreted as a direct result of mantle melting (e.g., Klein and Langmuir, 1987; McKenzie and Bickle, 1988). The evolved chemistry of MORB and rarity of very primitive lavas indicate that nearly all lavas erupted at the ridge crests are processed in magma chambers. However, whether fractionation is solely responsible for magma chemistry remains unquantified. Recent results from fast- and slow-spreading ridges (e.g., Rubin and Sinton, 2007; Lissenberg and Dick, 2008; Suhr et al., 2008; Godard et al., 2009; Drouin et al., 2009, 2010) indicate that significant reactions can occur between melts and lower crustal cumulates or mantle rocks. The extent to which melt-rock interactions bias our current understanding of mantle melting processes cannot be assessed without studying the genetically conjugate cumulate rocks with their daughter extrusive lavas (and ultimately the source mantle rocks). Eventually, what will be required is a bulk chemical inventory of a complete section of ocean crust.

The manner of passage of melt through the lower crust to the axial melt lens or to feed the dike and volcanic layers also remains poorly understood. Gabbros that crop out in ophiolites commonly exhibit fine-scale modal and geochemical layering, but these textures are difficult to reconcile with models of grain boundary flow of upwelling magma through a lower crust that mostly comprises a crystal mush (e.g., Korenaga and Kelemen, 1997). Discrete channels that feed magma into the axial melt lens or higher levels are yet to be identified in intact ocean crust (cf. MacLeod and Yaouancq, 2000).

The latent and specific heat from cooling and crystallizing magma is the principal driving force for hydrothermal circulation, with the energy available a function of the volume, distribution, and timing of magma intrusions. Within a few hundred meters of the ridge axis, the ocean crust appears completely solid to seismic waves and a clear Moho is generally observed. This requires that, at the very least, the latent heat of crystallization and sensible heat for cooling the magma to the solidus for the ~6 km of new crust at the ridge must have been exported from the system. The timescales are too short (<25,000 y) for this heat export to be achieved solely by conduction, requiring advection of heat by hydrothermal circulation. How this can be achieved in the upper crust is easy to envisage, but the importance and geometry of latent and sensible heat extraction from the deep crust by hydrothermal fluids remain poorly known and provide a key difference in competing models of magmatic accretion at fast-spreading ridges (Fig. F4) (Sleep, 1975; Henstock et al., 1993; Dunn et al., 2000; Garrido et al., 2001; Maclennan et al., 2005).

The compositions of fluids venting into the ocean at high-temperature black smokers and other types of vents are controlled by the physiochemical conditions and the extents of fluid-rock reactions within the crust (e.g., Mottl, 1983; Seyfried et al., 1999; Jupp and Schultz, 2000; Coumou et al., 2008). The rate of cooling of magma is in turn controlled by the extent of fracturing and resulting permeability, the consequent geometry and vigor of high- and low-temperature hydrothermal circulation, and the rates of fluid-rock exchanges. Some numerical models and ophiolite data (e.g., Maclennan et al., 2005; Bosch et al., 2004; Gregory and Taylor, 1981) require that seawater circulation extends to depths of several kilometers close to the ridge axis to mine the latent heat from deep in the crust and hence directly controls accretionary processes in the lower crust. Unfortunately, deep circulating fluid fluxes are poorly determined, and the conclusive geochemical tests of this scenario in an intact section of ocean crust remain to be conducted (e.g., Coogan et al., 2002, 2005; Van Tongeren et al., 2008). Sparse analyses of hydrothermal veins from gabbros indicate insufficient fluid volumes to significantly cool the lower crust (Coogan et al., 2007). The chemistry of black-smoker fluids suggests rock-dominated fluid exchange with the crust and regional recharge, but faults may play a role in facilitating the penetration of seawater-derived fluids to enable the cooling of the deep crust (e.g., Coogan et al., 2006). However, to date there is little evidence from intact ocean crust on whether faults, or other channels for seawater penetration down into the lower crust, are important for cooling the lower crust and for the advection of ocean-derived geochemical tracers or microbial populations to depth (e.g., Mason et al., 2010). Microbial populations seek out high thermal/chemical gradients; hence, the variation in the location/properties of faults and other zones of enhanced crustal fluid recharge are expected to determine the diversity of the ecosystem at depth within the crust.

An important recent advance comes from the recognition that the sheeted dike complexes of all intermediate to fast-spread systems studied (DSDP Hole 504B and ODP Hole 1256D and seafloor samples from Hess Deep and Pito Deep tectonic windows) provide relatively consistent estimates of axial high-temperature fluid fluxes (e.g., Teagle et al., 1998a, 2003; Gillis et al., 2005; Barker et al., 2008; Harris et al., 2008; Harris, 2011; Coggon, 2006; Nielsen et al., 2006; Chan et al., 2002). These estimates are all much lower than hydrothermal fluxes estimated from global seawater budgets, hydrothermal vent observations (e.g., Elderfield and Schultz, 1996), or studies of ophiolites (Bickle and Teagle, 1992), but their consistency with thermal calculations gives confidence in their validity. This sets the stage for estimates of chemical fluxes between this zone and the oceans and the impact of axial hydrothermal alteration on global chemical cycles (e.g., Davis et al., 2004; Vance et al., 2009).

Deep scientific ocean drilling is the only approach that can provide basic geologic observations on the formation and evolution of fast-spreading ocean crust

To date there remains a near-complete lack of direct observations regarding the accretion occurring beneath the dike layer at fast-spreading ridges. Importantly, we have well-developed but competing theoretical and geological models of the styles of magmatic accretion at fast-spreading ridges (Fig. F4). These models have been developed from a wide evidence base from marine geology and geophysics, as well as studies of ophiolites. Unfortunately, none of the best preserved ophiolites likely formed in major ocean basins (e.g., Miyashiro, 1973; Rautenschlein et al., 1985; Miyashita et al., 2003; Stern, 2004). Although ophiolite outcrops will continue to provide invaluable inspiration for ocean crustal studies, their direct relevance to intact ocean crust remains unproven. Although tests have been developed, the appropriate materials and observations to challenge these hypotheses remain elusive because the key processes of crustal accretion occur through magma intrusion deep within the crust. These critical samples and data can only be recovered by deep scientific drilling of intact ocean crust.

Top of page   |    Previous   |    Next