Seawater recharge into oceanic crust: IODP Exp 327 Site U1363 Grizzly Bare outcrop



[1] Systematic differences in sediment thermal and pore water chemical profiles from Integrated Ocean Drilling Program Site U1363 document mixing and reaction within the basaltic crust adjacent to Grizzly Bare outcrop, a site of hydrothermal recharge into 3.6 My-old basaltic crust. A transect of seven holes was drilled ~50 m to ~750 m away from the base of the outcrop. Temperatures at the sediment-basement interface increase from ~6°C to >30°C with increasing distance from the outcrop, and heat flow is suppressed within several hundred meters from the outcrop. Calculated fluid compositions at the sediment-basement interface are generally explained by mixing between bottom seawater and altered crustal basement fluids, with a composition similar but not identical to fluids from seeps at Baby Bare outcrop, located ~45 km to the northeast. Reactions within upper basement and overlying sediment affect a variety of ions (Mn, Fe, Mo, Si, PO43-, V, and U) and δ13DIC, indicating a diagenetic influence and diffusive exchange with overlying sediment pore waters. The apparent 14C age of basal pore fluids is much older than bottom seawater. Collectively, these results are consistent with seawater recharge at Grizzly Bare outcrop; however, there are strong gradients in fluid composition within 50 m of the outcrop, providing evidence for complex flow paths and vigorous mixing of young, recently recharged seawater with much older, more reacted basement fluid. The proximity of these altered fluids to the edge of the outcrop raises the possibility for fluid seepage from the outcrop in addition to seawater recharge.

1 Introduction

[2] The eastern flank of the Juan de Fuca Ridge has been the subject of numerous mapping, heat flow, sediment coring, seismic, and submersible expeditions, and targeted for three scientific drilling expeditions (Ocean Drilling Program (ODP) Leg 168 and Integrated Ocean Drilling Program (IODP) Exp 301 and 327) [e.g., Coggon et al., 2010; Cowen et al., 2003; Davis et al., 1992, and 1997; Fisher et al., 2003a, 2005, and 2011; Hutnak et al., 2006; Lever et al., 2010, Mottl et al., 1998, Orcutt et al., 2010, and Wheat et al., 2000 and 2010]. A central goal of these studies has been to elucidate coupled hydrogeologic, geochemical, and microbiological processes and properties within a ridge-flank setting, where fluid, heat, and solute fluxes are driven by lithospheric cooling and focused by basaltic basement roughness, crustal faulting and sediment accumulation [e.g., Fisher and Wheat, 2010].

[3] Many of the studies in this area have focused on fluid flow and related processes that occur on 3.5 to 3.6 My-old seafloor, as part of a regional, warm (up to ~65°C) ridge-flank hydrothermal system (RFHS). Seawater enters the volcanic oceanic crust through a seamount, flows laterally within the crustal rock below thick sediment, and emerges through another seamount after undergoing considerable heating, reaction, and diffusive exchange with overlying sediment pore waters [Wheat et al., 2000; Fisher et al., 2003a; Hutnak et al., 2006]. This is the only known RFHS for which the geometry of the primary fluid flow path, from recharge to discharge, is known with confidence.

[4] Considerable sampling has been accomplished at the discharge end of this system, with shallow gravity, piston and push coring and direct sampling of springs and boreholes, allowing the geochemical composition of crustal basement fluids to be resolved [e.g., Mottl et al., 1998; Wheat et al., 2002; Wheat et al., 2010]. Heat flow data collocated with seismic data and sediment core samples have allowed an assessment of thermal conditions, demonstrating near isothermality of upper basement temperatures, and a monotonic increase in seafloor heat flow, close to Baby Bare outcrop [e.g., Davis et al., 1992; Wheat et al., 2004a; Hutnak et al., 2006]. In contrast, much less is known about what happens during the first phases of water-rock-sediment interaction, soon after seawater enters the crust at a RFHS.

[5] On the basis of co-located thermal and seismic transects radiating from Grizzly Bare outcrop, Grizzly Bare outcrop is a site of seawater recharge to the crust [Fisher et al., 2003a; Hutnak et al., 2006]. One of these transects was positioned in the direction of a buried ridge that extends to the north, upon which Baby and Mama Bare outcrops reside and parallel to the present spreading center to the west. This transect was selected for measurements and coring on IODP Expedition 327 because the sediment-basement interface is relatively well imaged, thermal data from this transect are consistent with seawater recharge into the crust and it trends towards Baby Bare outcrop. Seven holes (in four locations) were drilled at Site U1363 along this transect. Thermal measurements were made and sediment was collected for pore water extraction and analysis to assess thermal and chemical changes in the upper basaltic basement. These data represents the first detailed assessment of upper basement conditions near a recharge site for a RFHS.

[6] Forty to fifty kilometers north of Site U1363, along the trend of a buried basement ridge, there are monotonic changes in fluid composition along an inferred fluid flow pathway even in heterogeneous permeable upper basaltic crust [Wheat et al., 2000; Hulme and Wheat, 2013]. We hoped to find similarly systematic changes in chemical and thermal conditions along the Site U1363 transect, allowing us to quantify rates of basement fluid alteration, elucidate processes related to crustal evolution, and assess microbial impact as cool oxic seawater evolves into a reducing, warm and altered crustal basement fluid. Results presented in this paper suggest that fluid flow, mixing, and reactions at Site U1363 are more complex than anticipated.

2 Regional Setting and Processes

[7] Seafloor spreading along the Juan de Fuca Ridge (Figure 1A), to the west of the field area, has created extensive bathymetric relief associated with faulting and the formation of abyssal hills. The volcanic oceanic crust is exposed close to the active spreading center, but sediment is generally thicker and more continuous towards the continental margin to the east where rapid sedimentation in this near-continental region has resulted in the burial of volcanic crust at an unusually young age [Underwood et al., 2005]. Sediment is typically 250–600 m thick where the seafloor is 3.5 to 3.6 My old, east of the active spreading center, but there are volcanic outcrops that penetrate through this thick sediment, remnants of off-axis volcanic events [Karsten et al., 1998; Becker et al., 2000] that now serve as entry and exit points for RFHS fluids. Sediment in this area is generally orders of magnitude less permeable than the upper basaltic crust [Spinelli et al., 2004; Becker et al., 2008; Fisher et al., 2008], limiting rapid flow of seawater into and hydrothermal fluids out of the crust to locations where highly permeable conduits such as basement outcrops or faults are exposed at the seafloor.

Figure 1.

The eastern flank of the Juan de Fuca Ridge has been the site of several drilling expeditions that include sediment and basalt recovery and borehole observatory installations (CORKs). Seven holes were drilled near Grizzly Bare outcrop on IODP Exp 327 (Site U1363).

[8] A cluster of three basaltic outcrops penetrates through sediment on 3.5 to 3.6 My-old seafloor 100 km east of the Juan de Fuca Ridge near 48°N: Papa Bare, Mama Bare, and Baby Bare outcrops [Davis et al., 1992; Mottl et al., 1998; Becker et al., 2000]. Baby Bare outcrop (Figure 1A) is the most extensively surveyed and smallest of the three features. Here, highly altered fluid discharges at a rate of 5–20 l/s and releases 2–3 MW of heat [Mottl et al., 1998; Wheat et al., 2004a]. Mama and Papa Bare outcrops are also known sites of hydrothermal discharge, although surveying of these seamounts is limited.

[9] No hydrothermal recharge sites have been identified on or near these three seamounts. However, basement fluids are thought to recharge through Grizzly Bare outcrop, 52 km to the south, based on thermal and geochemical data (Figure 1). Grizzly Bare outcrop is conical in shape, 3.5 km in diameter, and rises 450 m above the surrounding seafloor. Grizzly Bare outcrop was identified initially as a site of regional hydrothermal recharge based on systematic differences in basement fluid composition and patterns of seafloor heat flow immediately adjacent to the edifice [Wheat et al., 2000; Fisher et al., 2003a; Hutnak et al., 2006]. Seafloor heat flow is depressed within a few kilometers of the edge of basalt exposure along several transects [Fisher et al., 2003a; Hutnak et al., 2006]. Seismic reflection data allow determination of sediment thicknesses at locations where heat flow was measured. Downward continuation of seafloor heat flow data along these lines suggested basement temperatures of 60–65°C within several kilometers of the outcrop edge. Temperatures are much lower, and seafloor heat flow values are suppressed, closer to the outcrop edge, where sediment is 100–200 m thick. In contrast, warm altered fluids discharging from Baby Bare Outcrop cause a monotonic upward sweep of isotherms with proximity to the outcrop, resulting in extremely high seafloor heat flow adjacent to the volcanic outcrop [Davis et al., 1992; Fisher et al., 2003a; Wheat et al., 2004a].

[10] Regional scale fluid flow in upper volcanic crustal rocks from Grizzly Bare to Baby Bare outcrops is driven by a “hydrothermal siphon,” the difference in pressure at the base of more dense recharging (cool) and less dense discharging (warm) fluids. This is the same forcing that is likely to drive most sedimented ridge-flank hydrothermal discharge, helping to explain the global ridge-flank heat flow anomaly [Fisher and Harris, 2010; Fisher and Wheat, 2010]. However, hydrothermal circulation between Grizzly Bare and Baby Bare outcrops has virtually no regional influence on lithospheric heat loss [Davis et al., 1999; Fisher et al., 2003a; Hutnak et al., 2006] because the amount of fluid (and associated advective heat) that discharges from Baby Bare outcrop is relatively small.

3 Methods

[11] Procedures for sediment and basement coring during the expedition, and information about core recovery, are described in more detail elsewhere [Fisher et al., 2011]. Of the seven boreholes drilled at Site U1363, three were washed through the sediment section without coring to determine the thickness of the sediment prior to using the advanced piston coring (APC) system (Table 1). The APC system often provides higher quality core than does the extended core barrel (XCB) system, which is compatible with the same bottom hole assembly, but the XCB corer also allows collection of a small amount of upper basement rock, and potentially the sediment-basement interface itself. Neither core system is especially effective at recovering undisturbed core material from unlithified sand-rich lithologies, which tend to either be circulated away by the drilling and coring process, or become remobilized and settle out in the core liner. In general, the sediment cover in this area comprises hemipelagic mud (clayey silt to silty clay), thin-bedded turbidites (sand-silt-clay), and thick-bedded medium sand turbidites [Fisher et al. 2011].

Table 1. Summary of Borehole Configurations and Thermal Results
HoleSediment Thickness (m) aNtemp bCTR c (m2-K/W)Heat Flow (W/m2)Tsbi d (°C)
  1. a

    Sediment thickness determined from tagging basement with drill bit.

  2. b

    Number of successful sediment temperature determinations used to determine heat flow and sediment-basement interface temperature (along with bottom water temperature).

  3. c

    CTR = cumulative thermal resistance, calculated from empirical relation developed for Sites 1026 and 1027 in Davis et al. [1999].

  4. d

    Tsbi = temperature at the sediment-basement interface.


[12] Sediment pore waters were extracted by squeezing cored sediment that was processed within a nitrogen-filled glove-bag at room temperature [Fisher et al., 2011]. Pore water composition was determined using standard colorimetric, titration, inductive coupled plasma emission and mass spectrometery, and ion chromotraphy techniques and are tabulated [Fisher et al., 2011], except for the carbon isotopic measurements which were made on four 30- to 60-ml samples sealed in glass bottles and poisoned with mercuric chloride (Table 2). These carbon isotopic measurements were made at the National Ocean Sciences Accelerator Mass Spectrometry Facility from the dissolved inorganic carbon in the samples. Samples for methane gas concentrations were also collected immediately after core recovery from each location where a pore water whole-round core was collected; however, no detectable methane concentrations were recorded [Fisher et al. 2011]. Due to filter contamination issues, samples for nitrate concentrations measured shipboard were considered to be unreliable, and a separate pore water aliquot was collected on shore. These samples were collected from sediment whole-round cores that had been frozen at −80°C then subsequently thawed in a 4°C cold room. Pore water was extracted from these samples using acid-washed Rhizon samplers (Rhizosphere Research Products, Netherlands), and the samples were analyzed using the same analytical methods as used shipboard.

Table 2. Measured Carbon Isotopic Values and Calculated Values for the Uppermost Fluid in Basaltic Basement
  MeasuredMeasuredMeasuredBasementEndmemberc, dEndmemberc, e
  1. a

    Duplicate samples.

  2. b

    Two separately processed samples from the same core.

  3. c

    Assumes that the alkalinity and DIC are equivalent, given the pH of the pore waters.

  4. d

    Calculated assuming the measured 14C is a result of adding depleted C from the pore water to the value in the basement, thus 14Cmeasured Concmeasured = 14Cpore water Concpore water + 14Cbasement Concbasement.

  5. e

    Calculated assuming the measured δ13C is a result of adding −20‰. δ13C from the pore water to the value in the basement, thus δ13C measured Concmeasured = δ13Cpore water Concpore water + δ13Cbasement Concbasement.

  δ13C (‰)Δ14Cmmol/kgmmol/kg14C ageδ13C (‰)
2AaU1363G 3H6−13.1−792.692.912.6712,000−12
2BaU1363G 3H6−13.11−799.782.912.6712,000−12
3bU1363F 4H3−7.94−820.721.811.2010,000−1.8
4bU1363F 4H3−10.92−801.31.811.209,400−6.3
1U1363D 4X3−11.46−916.871.270.9818,000−8.9

[13] Ten determinations of in situ temperature were attempted using two tools, the APCT-3 tool and the SET probe [Fisher et al., 2011]. The APCT-3 tool is the third generation of an instrumented APC coring shoe [Heesemann et al., 2006]. An autonomous data logger, temperature sensor, and power supply are housed inside an annular cavity in a custom-designed APC coring shoe, and a modified core-catcher subassembly is attached above the shoe. The instrument is programmed to collect data at regular intervals (typically every 1 to 5 s). The instrumented coring shoe is held just above the mudline to collect bottom water temperature data, then lowered into place so that the core barrel seals the base of the drillstring, which is used as a hydraulic accumulator for piston coring. After the tool is fired into the formation, it is left in place for 6–7 min, to allow partial thermal equilibration following frictional heating resulting from insertion of the coring shoe. Data are recovered when the APCT-3 tool is returned to the surface, along with the associated piston core.

[14] The SET probe is a modified version of a conically tipped, push-in probe first deployed during ODP Leg 168 [Davis et al., 1997]. Like the APCT-3 tool, the SET probe is programmed prior to deployment to record data at a fixed time interval. The SET probe is deployed down the drill string in lieu of a conventional core barrel. It is latched into position with the tip of the tool extending ~1 m beyond the end of the bit, then pushed into the formation using the drill string, decoupled from the drill string, and allowed to record data for 6–7 min. The tool is pulled from the formation using the drill string and recovered by wireline. Because the diameter of the SET probe is only about 1 cm at the tip, its thermal response time is much faster than that of the APC coring shoe, meaning that a larger fraction of the thermal decay associated with probe insertion occurs during the limited measurement period.

[15] Data from both tools were processed using software that compares measured tool response to synthetic cooling curves based on tool geometry, tool properties, and formation properties. Processing is graphically interactive, which allows the user to identify irregularities in temperature time-records. Measured temperature values were compared to modeled tool response curves, then extrapolated to infinite time to estimate equilibrium temperatures. Each measurement was processed using a range of sediment thermal conductivity values, because Expedition 327 thermal conductivity data from sandy turbidites were highly scattered and often of low quality. Heat flow values were determined by combining calculated equilibrium sediment temperatures with a cumulative thermal resistance versus depth function, as determined for ODP Sites 1026 and 1027 based on seafloor heat flow, sediment thickness, and upper basement temperature data [Davis et al., 1999]. Comparison of lithologic and thermal conductivity data collected during ODP Leg 168 and IODP Expeditions 301 and 327 suggests that lithologies encountered during the three drilling expeditions were similar.

4 Results

[16] Sediment pore water chemical profiles are influenced by diagenetic reactions and exchange with basement fluids as observed at ODP Sites 1026 and 1027 and IODP Site U1301 [Davis et al., 1997; Fisher et al., 2005]. However, in contrast to these sites, pore waters from the base of the sediment column at IODP Site U1363 are cooler and typically less altered, with Mg concentrations of 36 to 28 mmol/kg (Figure 2). Sulfate reduction in the sediment column results in an increase in alkalinity, both showing an apex near the middle of the sediment column. At the base of the sediment column, sulfate and alkalinity values trend towards bottom seawater values because of exchange with a less altered basement fluid. Concentrations of Ca generally increase downhole, with the exception of a profile affected by carbonate precipitation, resulting from the high alkalinity produced by microbial sulfate reduction. Other elements, such as K and Rb, show changes in pore water concentrations with depth, ending near the sediment-basement contact at concentrations that are altered relative to seawater. Because of the K sampling artifact from squeezing [de Lange et al., 1992], we reduced the measured pore water concentration by 1.4 mmol/kg for all values in Figure 2.

Figure 2.

Selected pore water chemical depth profiles from IODP Site U1363 and seawater values. Pore waters are affected by sediment diagenetic processes and diffusive fluxes from the overlying seawater and underlying basaltic basement fluids. Measured data are in Fisher et al. [2011] and extrapolated basement fluids compositions are presented in Table 3. Because of the K sampling artifact from squeezing [de Lange et al., 1992], we reduced the measured pore water concentration by 1.4 mmol/kg.

Table 3. Composition of Fluids in Basaltic Basement at IODP Site U1363 and Baby Bare Outcrop Compared to the Composition of Bottom Seawater
SiteBottom aU1363G b, cU1363F bU1363B bU1363C/D bBaby a
  1. a

    Data are from Wheat and Mottl [2000], Wheat et al. [2002], and Sansone et al. [1998].

  2. b

    Data are from Fisher et al., 2011.

  3. c

    Samples from deeper than 17.5 m result from the piston sucking in and homogenizing basal sediment without further penetration even though depths are recorded as deeper than 17.5 m.

  4. d

    [Walker et al., 2007].

 Seawater    Bare
Latitude (N)---47º17.312’47º17.326’47º17.352’47º17.574’47º42.37’
Longitude (W)---128º2.170’128º2.137’128º2.106’128º2.762’127º47.15’
Distance From Outcrop (m)---5011517777052,000
Sediment Thickness (m)---17.535.057.0231.20
Deepest Pore Water (m)---17.532.253.3222.7---
Basement Temp (°C)1.867123364
Mg (mmol/kg)52.636353528.60.98
Chlorinity (mmol/kg)542544536546551554
Sulfate (mmol/kg)
Alkalinity (mmol/kg)2.52.671.201.710.980.43
Br (mmol/kg)---0.820.800.790.81---
Na (mmol/kg)467473464474471473
Na/Cl (mol/mol)0.8610.8690.8650.8690.8540.853
Ca (mmol/kg)10.321.022.623.335.055.2
Sr (mmol/kg)86858390101110
Ba (mmol/kg)0.150.350.070.591.040.43
K (mmol/kg)
Li (mmol/kg)26.66217.918.216.79.0
Rb (mmol/kg)1.371.411.71.330.531.12
Cs (nmol/kg)2.2233.12.25.3
Si (µmol/kg)190420300310350360
B (µmol/kg)410460580450270570
Nitrate (µmol/kg)39.22.52120.8
Ammonium (µmol/kg)0.310711017022076
Phosphate (µmol/kg)
Mn (µmol/kg)0.001556791002.9
Fe (µmol/kg)0.0012.7003<0.05
V (nmol/kg)38.46146510
U (nmol/kg)
Mo (nmol/kg)100870560550280297
δ 13C−0.6−12−4--−8.9−0.6
14C Age (Years)230012,00010,000--18,00012,000 d

[17] The composition of the basement fluid was calculated for most holes by linear extrapolation of the deepest four to ten samples to the depth of uppermost basement (Holes U1363B, D, and F) when gradients exist or by averaging the deepest four samples (Table 3). This analysis is adequate for most chemical species, but not for several trace elements (Pb, Cr, Co, Cu, and Zn) [e.g., Wheat et al., 2004b]. The number of samples used for these extrapolations differed from hole to hole and with ionic species, based on the reactivity of each ion and the shape of the associated profile. In the case of Hole U1363G, the deepest sample was collected right at the sediment-basement interface, so the composition of this fluid is considered to be that of the upper basement fluid. At this hole, samples were collected at a “curation” depth greater than 18 m. These “deeper” samples are interpreted to be a sediment slurry that was sucked into the core liner during the APC coring process (as was apparent from visual inspection of cores after recovery). Interestingly, the compositions of these slurry samples match those from the sample collected at 17.5 m depth and basal pore water chemical trends at the sediment-basement interface.

[18] In situ sediment temperatures were determined in Holes U1363B, U1363C, U1363F, and U1363G (Table 1, Figure 3). Two sediment temperatures plus a bottom-water determination were used to assess the thermal gradient and seafloor heat flow for these holes, except for Hole U1363G, which was only 17.5 m deep, allowing for a single sediment temperature determination just above basement. Sediment temperatures were particularly well determined at the base of Holes U1363F and U1363G, where measurements were made adjacent to the basement contact. Data from Holes U1363B and U1363C were not collected as close to the basement contact, but the depth to basement is known with confidence, and the consistency of measured thermal gradients allows confident extrapolation (Figure 4; Table 1).

Figure 3.

Thermal data from IODP Site U1363. Interpreted temperature and heat flow values are listed in Table 1.

Figure 4.

Reflection seismic and thermal data adjacent to Grizzly Bare outcrop and Site U1363. Location of transect is shown in Figure 1B. A. Seismic data from Zühlsdorf et al. [2005] and Hutnak et al. [2006], with isotherms estimated from downward continuation of heat flow data. B. Seafloor heat flow data (small squares) from Hutnak et al. [2006]. Large squares are borehole data reported in this study. C. Extrapolated temperatures at the sediment-basement interface from downward continuation of the seafloor and borehole heat flow data.

[19] Heat flow determinations from Holes U1363C and U1363F are similar to values determined from nearby surface probe measurements, and the value determined for Hole U1363G is similar to surface probe data from a position 65 m away (Figure 4B). The heat flow determined for Hole U1363B is about 50% greater than that determined during the surface probe survey [Hutnak et al., 2006]. Basement temperatures estimated from extrapolation of borehole thermal data are generally similar to values calculated from surface probe thermal data and estimated sediment thickness values from seismic reflection data (Figure 4C). The consistency of these results is impressive considering the difficulty in imaging the sediment-basement interface from seismic reflection data along a sloped surface adjacent to Grizzly Bare outcrop. The key result from these thermal studies is confirmation that upper basement temperatures are suppressed within several hundred meters of the edge of Grizzly Bare outcrop [Figure 4, Fisher et al., 2003a, Hutnak et al., 2006].

5 Discussion

5.1 Thermal Data: General Hydrologic Pattern

[20] The prevailing paradigm of RFHS is that seawater warms as it enters the basement, reacts with volcanic crustal rocks, is affected by microbial processes, and exchanges with overlying sediment pore waters. Thus, the temperature of the basement fluid is affected by transport processes, which include advection and conduction, whereas the fluid composition is affected by advection, diffusion, and reaction.

[21] Fisher et al. [2003a] and Hutnak et al. [2006] show heat flow profiles adjacent to both recharging and discharging outcrops, and compare values observed to those inferred from conductive and coupled (fluid-heat) numerical models. Hutnak et al. [2006] also show calculated temperatures along the sediment-basement interface, as inferred from downward continuation of seafloor heat flow data. In areas of observed hydrothermal discharge, heat flow values increase monotonically with increasing proximity to the exposed basement outcrop, and temperatures at the sediment-basement interface remain elevated, well above bottom water values [e.g., Baby Bare outcrop, Figure 10 of Hutnak et al., 2006].

[22] Transects of thermal profiles from areas of inferred recharge have a contrasting pattern. In locations with the most vigorous recharge, heat flow is suppressed within ~500 m of the outcrop edge, and temperatures at the sediment-basement interface decrease towards bottom seawater temperature as the outcrop is approached [e.g., Grizzly Bare outcrop, Figure 12 of Hutnak et al., 2006]. However, the monotonic reduction of heat flow may not extend all the way to the outcrop edge, for two reasons. First, conductive refraction tends to generate a local, positive heat flow anomaly at the break in slope between the flat seafloor and the sloped basalt edifice. In addition, secondary hydrothermal convection within the outcrop generates local anomalies [elevated and suppressed values, e.g., Figure 22 of Hutnak et al., 2006; Kawada et al., 2011]. Some thermal transects around Grizzly Bare outcrop show a mixed thermal pattern, with small local anomalies superimposed on a broader pattern consistent with recharge [Figures 12–14 from Hutnak et al., 2006]; one of the transects radiating from Grizzly Bare outcrop, to the northwest of the Site U1363 transect, shows evidence of discharge [Figure 15 from Hutnak et al., 2006].

[23] The thermal transect selected for placement of the Site U1363 transect, and verified with boreholes measurement, shows a clear recharge signature: a zone of low heat flow located several hundred meters from the outcrop edge, and sediment-basement temperatures that decrease towards bottom seawater values close to the outcrop. The monotonically elevated heat flow values seen adjacent to Baby Bare outcrop, where vigorous discharge has been confirmed, are clearly absent [Fisher et al., 2003a; Hutnak et al., 2006; Wheat et al., 2004a]. However, there are three elevated heat flow values close to the edge of Grizzly Bare outcrop along this transect. There is no evidence of upward flow through sediment (shallow thermal gradients in sediment are clearly conductive), and heat flow is suppressed farther from the outcrop (Figure 4). As described earlier, conductive refraction at the break in slope and secondary convection in basement could both contribute to the near-outcrop anomaly, which is superimposed on the broader thermal indications of recharge. The Site U1363 transect is located between transects showing an even stronger recharge signature (to the southeast), and one showing a modest discharge signature (to the northwest) [Fisher et al., 2003a, Hutnak et al., 2006]. Site U1363 appears to be located within the transition between these regimes.

5.2 Chemical Data: Transport and Mixing Close to Grizzly Bare Outcrop

[24] Although thermal and chemical data are broadly indicative of RFHS recharge through Grizzly Bare outcrop [Wheat et al., 2000; Fisher et al., 2003a; Hutnak et al., 2006], data presented in the present study suggest complexity in flow paths, including recharge, fluid flow, and mixing in and out of the plane of the Site U1363 transect. If there were conservative mixing between recharging and reacted hydrothermal fluids along this transect, one might expect to observe a linear trend in a cross plot of temperature and Mg concentration at the sediment-basement interface. End-members would be defined by bottom seawater (cold, high Mg) and altered hydrothermal basement fluid (warm, low Mg), similar in composition to fluids sampled from Baby Bare springs and nearby instrumented boreholes (Table 3). A linear mixing trend such as this was anticipated along the Site U1363 transect because laboratory studies of seawater-basalt experiments at low temperatures indicate very slow reaction rates [Seyfried, 1977], consistent with observations from a variety of settings around the globe [Elderfield et al., 1999; Fisher and Wheat, 2010]. Data from Holes U1363C/D (located ~750 m from the outcrop edge) lie on a mixing trend of this kind (Figure 5), consistent with advective transport and conservative mixing. However, data from the other holes (U1363B, F, and G) closer to the outcrop fall below this linear mixing trend (Figure 5), indicating that either there is a preferential loss of either heat or Mg. We see similar trends when plotting temperature versus other ions (e.g., Ca, alkalinity, and K).

Figure 5.

Temperature and magnesium concentration for selected fluids. The green squares represent the calculated values for the upper basaltic basement fluids from Site U1363 (Table 3). Black circles are similar data from ODP Leg 168 [Elderfield et al., 1999]. The red and blue data are experimental data after about 1 year of reaction [Seyfried, 1977; Seyfried and Bischoff, 1979]. The yellow line represents conservative mixing of bottom seawater with spring fluids from Baby Bare outcrop (Table 3). The arrow highlights the loss of heat in fluids nearer Grizzly Bare outcrop and at ODP Sites 1030/31.

[25] The non-conservative loss of Mg (and other solutes) is unlikely in this setting, based on experimental and environmental data at the observed temperatures [Seyfried, 1977; Fisher and Wheat, 2010]. For example, old altered basement fluids at ODP Site 1023 (~15°C) are less altered (47. 2 mmol Mg/kg) than the cooler boreholes at Site U1363, which should experience less alteration because at lower temperatures reactions are more sluggish. In contrast, heat loss is more likely because of the very high thermal diffusivity of crustal rocks in comparison to solute diffusivity. As a result, a crustal region adjacent to a recharge channel can cool without significant geochemical modification. A similar decoupling of heat content and fluid composition in basaltic crust on ridge flanks has been documented in boreholes as they return to their natural state after being perturbed by drilling operations [Wheat et al., 2003; Wheat et al., 2010], and as altered fluids rise from deeper in the crust towards the seafloor [e.g., ODP Sites 1030 and 1031; Elderfield et al., 1999].

[26] Our interpretation of Site U1363 thermal and geochemical data is that recharging seawater enters the crust in the vicinity of Site U1363, reducing upper crustal temperatures, but mixing, reactions and fluid flow in and out of the plane of the drilling transect and from below lead to complex geochemical and thermal patterns at the sediment-basement interface. Further from the outcrop (Holes U1363C/D) vigorous local convection and mixing result in a fluid composition consistent with simple mixing trends. It should not be surprising that fluid pathways are complex in a system such as this one, particularly considering the heterogeneity in upper crustal stratigraphy and structure in this region. This interpretation is consistent with variations seen in basement fluid samples from ODP Sites 1026 and 1027, ~45 km to the north [Davis et al., 1997]. Although these sites are just 2.2 km apart, basement fluids from these sites are chemically distinct, suggesting a lack of mixing between them. Contrasts in fluid temperatures and compositions within <1 km of Grizzly Bare outcrop are similarly profound with oxic seawater in close proximity to warm reduced basement fluids. The potential for strong vertical and lateral gradients in redox and other geochemical characteristics in basement provide conditions conducive to microbial activity, similar to strong redox boundaries that are commonly found at the sediment-basement interface [Engelen et al., 2008].

5.3 Extent of Basement and Sediment Reactions

[27] The extent of reaction within basaltic basement is assessed by comparing the fluid composition in upper basement to a mixing trend defined by regionally warm and altered basement fluids (e.g., having a chemical composition similar to springs from Baby Bare outcrop) and bottom seawater. As discussed above, if conservative mixing processes are dominant, then data for a chemical species plotted versus Mg would lie on a mixing line, given the slow rate of reaction for Mg at temperatures <25°C. Such trends are evident for some ions (Figure 6), but others fall off a simple mixing trend, indicating that reactive processes and/or diffusive exchange with overlying pore waters are important in controlling fluid concentrations. Probable reactions that take place are assessed by comparing the results with several data sets: (1) water-rock experiments [Seyfried, 1977; Seyfried and Bischoff, 1979], (2) estimated fluid compositions for upper basement at ODP Sites 1023–1025 with temperatures (15–39°C) that overlap those from IODP Site U1363 and ODP Sites 1028 and 1029 [Elderfield et al., 1999], (3) fluids collected following restoration of ambient crustal conditions in IODP Hole U1301A [Wheat et al., 2010], and (4) basal sediment pore water gradients that provide a measure of exchange with upper basement fluids [e.g., Elderfield et al., 1999; Wheat et al., 2000; Hulme and Wheat, 2013].

Figure 6.

Dissolve ion versus magnesium concentrations for selected fluids. The green squares represent the calculated data for the upper basaltic basement fluids from Site U1363 (Table 3). Black circles are similar data from ODP Leg 168 [Elderfield et al., 1999]. The red and blue data are experimental data after about 1 year of reaction [Seyfried, 1977; Seyfried and Bischoff, 1979]. The yellow line represents conservative mixing of bottom seawater with spring fluids from Baby Bare outcrop (Table 3). These data highlight mixing and reaction within the basaltic crust and diffusional inputs from overlying sediment pore waters.

[28] The Ca-Mg data from Site U1363 generally lie on the mixing line between reacted altered basement fluids and seawater (Figure 6). No such mixing scenario is suggested for experimental data or data from elsewhere along the ODP Leg 168 drilling transect, for which the primary influence on composition seems to be reaction temperature. Along the Leg 168 drilling transect, the removal of Mg from seawater is balanced by a crustal Ca source, maintaining charge balance, with the extent of reaction determined mainly by the temperature of reaction [e.g., Mottl and Wheat, 1994; Elderfield et al., 1999; Fisher and Wheat, 2010]. The fluid residence time must also be important, given that neither the 25°C nor the 70°C experiments [Seyfried, 1977], which were conducted for 1.5 and 0.5 years, respectively, reached concentrations equal to those from borehole samples at similar temperatures (Figure 5). The 14C ages of samples from the ODP Leg 168 transect are thousands of years old [Elderfield et al., 1999]. Even if the actual fluid “age” is considerably younger [e.g., Stein and Fisher, 2003], sufficient time has passed for these fluids to react with basalt, implying these fluids are “equilibrated” with basalt and with diffusive fluxes of Mg and Ca to/from overlying sediment pore fluids. In other words, the diffusive exchange with overlying sediment pore waters for these ions appears insignificant relative to rates of reaction [Wheat et al., 2000; Hulme and Wheat, 2013].

[29] A cross plot of alkalinity and Mg data from the Site U1363 transect also places data close to a mixing line, suggesting that mixing near the outcrop influences alkalinity (Figure 6). However, there appears to be some removal of alkalinity (data falling below the mixing line), consistent with experimental data that show depletions in CO2. In contrast, Mg concentrations are higher (relatively less reacted), indicating that reactions removing alkalinity are more rapid than those removing Mg.

[30] Similarly, K-Mg data from Site U1363 holes close to the outcrop (T <15°C) fall along the mixing line, suggesting minimal reaction relative to mixing. Experimental data are consistent with this pattern. However, at Hole U1363C/D, where the upper basement temperature is ~33°C, K is removed relative to Mg, consistent with observations from the high-temperature end of the ODP Leg 168 transect, and perhaps indicating a greater sediment influence. Site U1363 Li-Mg data (not shown) also fall on a mixing line, with the exception of data from Hole U1363G, which has a higher Li concentration relative to the trend. Although, Li tends to be removed from fluids by reaction with basaltic crust at these cool temperatures, the diffusive flux of Li from the overlying sediment tends to increase the concentration in basement fluids with time [Wheat et al., 2000]. Thus, the longer the fluids reside in basement, the larger the influence from sediment and the higher the Li concentration.

[31] Sulfate concentrations in RFHS can be explained by the integrated diffusive loss to the overlying pore waters with minimal removal in basaltic basement by microbial processes [Elderfield et al., 1999; Wheat et al., 2000; Lever et al., 2010; Hulme and Wheat, 2013]. For example, there is a monotonic decrease in sulfate concentrations along the transect from Baby Bare outcrop to the north side of Mama Bare outcrop as a result of continuous diffusive loss to overlying sediment pore waters along a flow path [Wheat et al., 2000; Hulme and Wheat, 2013]. Because of this diffusive control, a plot of SO4 versus Mg data from the ODP Leg 168 transect does not form a linear mixing trend. Most of the Site U1363 data lie above the Mg-SO4 mixing line (Figure 6). This trend suggests that the altered seawater from Hole U1363 C/D is less influenced by diffusive loss of sulfate to the overlying sediment than that expected between mixing of reacted basement fluid and seawater. Thus, these warm altered fluids must be sourced from a location where the basement fluid is relatively evolved with respect to thermally controlled elements (e.g., loss of Mg and gain of Ca), but less affected by sedimentary diffusive exchange relative to fluids that seep from Baby Bare outcrop. This suggests that these fluids may have originated at least several kilometers from Hole U1363 C/D where basement temperature are ~65°C, or may have undergone reaction deeper in the crust [e.g., ODP Sites 1030 and 1031; Elderfield et al., 1999]. Reactions at greater depth cannot be the sole explanation for the chemical data, because sulfate values are depressed relative to bottom seawater, requiring at least some sulfate removal by exchange with overlying sediment.

[32] Mn-Mg data from Site U1363 fall well above the mixing line, as do the other environmental data and the 70°C experimental data (Figure 6). This indicates that Mn is one of the more reactive elements within basaltic crust; however, it seems likely that elevated Mn values from Site U1363 are also influenced by exchange with sediment pore waters. Likewise, estimated concentrations of Mo and ammonium (and Si and Fe; not shown) in upper basement fluids appear to be influenced by sediment pore waters.

[33] One of the goals for sampling close to the sediment-basement interface at Site U1363 was to document variations in the redox state in the basement fluid with distance from the outcrop. Nitrate is a particularly sensitive redox indicator, and steep nitrate gradients were detected near the seafloor (in the upper 20 cm) in pore fluids recovered while gravity coring in this area (C.G. Wheat and M. Mottl, unpublished data). Unfortunately, we were unable to document steep nitrate gradients in basal sediment recovered from Site U1363, perhaps because we did not recover high-quality samples from close enough to the sediment-basement interface, or because there is so little nitrate in upper basement fluids. If the latter interpretation were correct, then nitrate is consumed very rapidly in upper basement fluids in this area. Likewise phosphate, V, and U (not shown) in basement fluids appear to be consumed rapidly along the Site U1363 transect (Figure 6).

[34] By examining all of these relationships, experimental data, and data from other field sites, we get a better understanding of which elements are most reactive in this setting. The sulfate data indicate that altered fluids from Site U1363 show less evidence for diffusive loss than do fluids that seep from Baby Bare springs. Furthermore, there is evidence that those ions that are greatly affected by diffusive exchange show more alteration closer to Grizzly Bare outcrop, perhaps indicating a greater influence of sediment sources (e.g., Mo, Si, Rb, Li, and sulfate: Table 3). This interpretation requires some movement of reacted basement fluids across the drill sites from warmer and/or deeper parts of the crust towards the outcrop (Figure 7). Thermal data suggest that there may be discharge of hydrothermal fluids along one of five transects measured adjacent to Grizzly Bare outcrop [Figure 15 in Hutnak et al., 2006], but the amount and extent of discharge close to the Site U1363 transect must be modest or else there would be less suppression of upper basement temperatures. More generally, these result emphasize the complexity of crustal flow paths around Grizzly Bare outcrop, with newly recharged and older, more reacted fluids maintaining distinct geochemical characteristics and gradients, even as flow occurs from recharge to discharge sites.

Figure 7.

Conceptual interpretation of regional thermal and geochemical data from Site U1363, and results of analytical and numerical studies [e.g., Wheat et al., 2000; Fisher et al., 2003a; Hutnak et al., 2006]. The seismic reflection profile and isotherms are the same as shown in Figure 4A. Mg and Ca concentrations (mmol/kg) and approximate upper basement temperatures (°C) are shown in boxes above drill sites. Similar parameters are also shown for recharging bottom seawater (large blue arrow). Cold recharging fluid and ambient (reacted, warm) basement fluids mix and react, and there is diffusive exchange for many solutes with the overlying sediment. The most intense recharge appears to occur off the Site U1363 transect, but it is close enough to impart a recharge signature. Inset shows a cartoon of a hypothetical, three-dimensional flow system, intended to place the Site U1363 drilling transect in context. There are likely to be a small number of recharge and possible discharge sites distributed irregularly around the perimeter of Grizzly Bare outcrop. The Site U1363 transect must be relatively close to a recharge zone because of the observed suppression of basement temperatures near the outcrop, but the fluid composition indicates that a component of flow in and out of the transect profile, and/or exchange with fluids from greater depth must occur. Some of the fluid recharging at Grizzly Bare outcrop is thought to discharge at Baby Bare outcrop, 52 km to the north [Wheat et al., 2000; Fisher et al., 2003a].

5.4 Residence Time and Apparent 14C Ages

[35] Four 30 to 60 ml pore water samples from near the sediment-basement interface were analyzed for their 14C age (Table 2). As with the other chemical data, we used these data to extrapolate to the 14C age of the fluid in upper basaltic basement, assuming that the alkalinity is a measure of DIC dilution of the end-member with the added C from depleted 14C from basal sediment. Although the 14C of individual samples shows a monotonic change (older) away from the outcrop, when the values are corrected for dilution with diagenetic DIC the monotonic trend vanishes. Surprising, the 14C age of these fluids is tens of thousands of years, similar to values from the ODP sites and boreholes [Elderfield et al., 1999; Walker et al., 2007].

[36] The δ 13C for the DIC in upper basaltic basement was similarly calculated (Table 2). We assumed a diagenetic δ 13C of −20‰, consistent with a plankton/sediment source [McCorkle et al., 1985; Rau et al., 2001; Zeebe, 2007]. The calculated values range from −1.8‰ to −12‰ without a systematic trend with distance from the outcrop. These values span those measured by Sansone et al. [1998], who analyzed spring fluids from Baby Bare outcrop reporting near seawater δ 13C values (~ −0.6‰) that are slightly different than seawater values measured reported by Walker et al., 2007; −1.3 and −1.5‰. Walker et al., 2007 analyzed borehole fluids from ODP Hole 1026B and harpoon fluids from Baby Bare outcrop reporting δ 13C values of −5.8 to −9.7‰, that were interpreted to reflect systematic carbonate precipitation with input of basaltic CO2. Spring fluids from Baby Bare outcrop were devoid of sediment artifacts (e.g., no excess ammonium and dissolved silica), unlike borehole fluids from Hole 1026B [Wheat et al., 2004b]. Data to determine if sediment influences could affect the Baby Bare harpoon δ 13C data were not provided [Walker et al., 2007]. An alternative explanation for these depleted δ 13C values is a sediment source, stemming from diffusive exchange with pore waters. Samples from Site U1363 are affected by sedimentary carbon and when corrected for values in basement fluids they are less depleted than the measured values. Note that one datum from Site U1363 is not consistent with Rayleigh fractionation with reasonable amounts of basaltic C inputs.

[37] Nevertheless, the carbon isotopic story is complex and does not follow those elements that are controlled by diffusive exchange with overlying pore water (e.g., sulfate) or reaction with basalt (e.g., Mg). The complete data set and in particular the two types of isotopic C data, suggest a complex mixing scenario near a recharge area for a RFHS, yet the isotopic DIC data as a whole are inconsistent, possibly resulting from sediment artifacts. If there are not artifacts, the DIC data are similar to other ions that are thought to be highly reactive in this setting. This conclusion is not anticipated because pore water alkalinity gradients near the sediment-basement interface are much smaller than those of other ions. Also, carbonate precipitation is ubiquitous and relatively quick in this environment where Ca, alkalinity, and temperature variations are systematic. A highly reactive DIC pool is not consistent with systematic alkalinity values nor microbial rates [Lever et al., 2010]: additional studies are required to sort out the DIC story.

6 Summary and Conclusions

[38] We present new pore water chemical and thermal data from IODP Site U1363 at the base of Grizzly Bare outcrop, a site where there is evidence for seawater recharge into the crust as part of a regional RFHS. A transect of boreholes extending radially from the outcrop reveals a mixture of seawater and thermally altered fluids, similar in composition to warm (64°C) hydrothermal fluids sampled ~50 km to the northeast. Surprisingly, only 50 m from exposed basalt on Grizzly Bare outcrop, this mixture indicates 66% (recharged) seawater and 34% thermally altered fluids (based on Mg data). This mixture decreases to about 50:50 at a distance of ~750 m from the basaltic outcrop. However, the temperatures of these fluids are much lower that one would find based on conservative mixing. This requires the conductive loss of heat from these fluid mixtures during transport. Not all of the measured ions are consistent with the ratios cited above, because of non-conservative exchange or reactions.

[39] Results from the Site U1363 transect do not indicate a simple pattern of monotonic change with distance from the outcrop. Instead, we see evidence for complex flow paths, mixing of distinct water types, and reaction. The data indicate a strong sedimentary signature close to the outcrop, consistent with a potential flow path towards the outcrop even though the thermal budget clearly requires nearby seawater recharge. There was no evidence for fluid seepage within thin sediment around Grizzly Bare outcrop during a site survey expedition [Zühlsdorff et al., 2005; Hutnak et al., 2006; Wheat and Mottl, unpublished data], but this is a large feature, and it is possible that diluted altered basement fluids exit the seafloor somewhere on or near the edifice, consistent with thermal data that show evidence for complex fluid flow patterns adjacent to the edifice.

[40] Results from this study illustrate how challenging it can be to “map” hydrologic flow paths in the ocean crust. Even around seamounts and other basement outcrops where inflow or outflow are known to occur, the primary channels through which fluids, heat, and solutes move through the crust remain poorly constrained by sampling at a small number of locations. Resolving these flow pathways is important, because they can result in large gradients in redox-sensitive solutes (anoxic, altered, and warm basement fluids relative to cold, oxic bottom seawater) and thus have significant implications for the evolution of the oceanic lithosphere and subseafloor microbial activity and ecology. The potential for strong vertical and lateral gradients in redox and other geochemical characteristics in basement provides conditions conducive to microbial activity, similar to strong redox boundaries that are commonly found at the sediment-basement interface [Engelen et al., 2008]. The Site U1363 transect is the first to systematically sample close to a known site of hydrothermal recharge on a RFHS. Resulting samples and data suggest that careful site selection and extensive spatial coverage will be required to resolve the complexity and implications of recharge guided by seamounts for linked thermal-chemical-microbial-geological processes.

7 Acknowledgments

[41] This work was made possible through the Integrated Ocean Drilling Program and the dedicated personnel that make it possible to collect quality deep-sea sediment cores. Shore-based funding was provided from the U.S. Science Support Program and the Center for Dark Energy Biosphere Investigations (C-DEBI). This work was supported by grants from the U.S. National Science Foundation, OCE 0939564 (with subawards to CGW and ATF), OCE-1030061 (CGW), and OCE-1031808 (ATF) and grants from the U.S. Science Support Program (CGW, SMH, ATF) and the Danish National Research Foundation and the Max Planck Society (BNO). This is C-DEBI contribution i57.