IODP

doi:10.2204/iodp.pr.345.2014

Site summary

Site U1415 is located along the southern slope of the intrarift ridge between 4675 and 4850 mbsl (Fig. F6). A series of 16 holes were drilled at Site U1415, two ~110 m below seafloor (mbsf) reentry holes (U1415J and U1415P), five single-bit holes (U1415E and U1415G–U1415I), two failed reentry holes (U1415K and U1415M), and six holes in which jet-in tests were conducted to assess sediment thickness (Holes U1415A–U1415D, U1415F, and U1415L) (Fig. F8, F9; Table T1).

Site U1415 is centered on a ~200 m wide, flat-lying east-west–trending bench at 4850 mbsl that is covered with approximately <10–30 m of largely gabbroic rubble overlain by a few meters of pelagic sediment mixed with lithic debris. Holes U1415E–U1415J are in an area where primitive (Mg# = 75–85) gabbroic rocks were recovered from above and below the bench during the JC21 site survey (see distribution of olivine gabbros in Fig. F7; see also Fig. F8). Holes U1415K–U1315N are situated along a ~100 m wide, flat-lying shoulder at ~4675 mbsl, ~160 m shallower than the bench (Fig. F8), where slightly more evolved gabbroic rocks were recovered during the JC21 site survey (Fig. F7).

Specific hole locations were selected in the general area of the proposed drill sites (HD-01B to HD-03B) using a combination of geomorphology, seafloor observations, and shallow subsurface seismic data. Early in the expedition, exploratory visual and seismic surveys, combined with jet-in tests, showed that the sediment thickness on the bench was at most a few meters. This meant deployment of a standard reentry system was not feasible; thus a hard rock reentry system–style nested free-fall funnel with casing approach was used (see “Operations summary”). Our early drilling results lead us to reevaluate the origin of the seafloor morphology observed in the microbathymetry of the bench area. Small-scale topography was reinterpreted to be related to mass wasting above and within the bench rather than extensional faulting along the bench, as previously thought. Thus, the strategy for placing the drill holes was modified slightly to focus on structural promontories separating areas where slumping may have dominated.

Drilling at Site U1415 also revealed that, at least locally, the bench was not a coherent fault block as originally proposed. Near-bottom 3.5 kHz seismic data suggest the presence several fault or slump blocks with a complicated structural relationship along and across strike. Variations in inclination of stable, high unblocking temperature magnetic components indicated that at least two blocks had been sampled in each hole, providing evidence of relative displacements of individual, internally coherent units. The scale of these blocks can be assessed in the two deepest holes, U1415J and U1415P (Figs. F11, F12). In Hole U1415J, consistency in the dip of magmatic foliation and magnetic inclinations identify two discrete blocks with vertical thicknesses of 29 and ≥50 m. In Hole U1415P, rock type and magnetic inclinations suggest two blocks of 65 and >42 m thickness. The similarity between the Troctolitic Series (see below) recovered in Holes U1415J and U1415P suggests that the blocks could originate from comparable stratigraphic levels.

The evidence of local mass wasting observed at Site U1415 and the prediction that mass wasting is a regional phenomenon (Ferrini et al., submitted) call into question the sensibility of assigning a stratigraphic position for any samples recovered along the southern slope of the intrarift ridge. The exception is the crest of the intrarift ridge. Here, the stratigraphic position of the gabbroic rocks is well constrained as being shallow gabbros, close to the sheeted dike–gabbro transition. This conclusion is based on presence of dolerites with EPR compositions interpreted as sheeted dikes along the northern slope of the intrarift ridge (Fig. F6) and cooling rates comparable to the uppermost gabbros along the Northern Escarpment of the Hess Deep Rift (Coogan et al., 2007). Thus, we do not present our expedition results in a depth context and strongly caution against the use of depth profiles for any samples recovered from the southern slope of the intrarift ridge (cf. Hékinian et al., 1993; Lissenberg et al., 2013).

A precise stratigraphic context for Site U1415 is not required to address most of the specific scientific objectives of Expedition 345. It is important, however, to comment on what can be constrained concerning the original depth of the Site U1415 primitive gabbros in the EPR plutonic crust. No primitive gabbros occur within 1 km of the base of the sheeted dike complex along the northern escarpment of Hess Deep Rift. Primitive olivine gabbros along the southern slope of the intrarift ridge first crop out at 3800–4000 mbsl, within 2.0 km of the inferred sheeted dike complex (Lissenberg et al., 2013) (Fig. F6). As these samples were recovered along a very steep slope inferred to be the head wall of a large mass wasting feature, we suggest they are in place in the context of the intrarift ridge block itself (unlike samples recovered at greater depths). Thus, we conclude that the primitive lithologies at Site U1415 formed at a minimum of 2 km beneath the sheeted dike complex, in the lower half to one-third of the 3.8–4.8 km thick plutonic sequence.

Magmatic accretion of fast-spreading East Pacific Rise lower crust

Rock types and units

At Site U1415, igneous rocks were recovered from massive blocks and overlying surficial rubble and lithic sands and gravels. Olivine gabbro and troctolite are the major plutonic rock types at Site U1415, with minor gabbro, clinopyroxene oikocryst-bearing troctolite, clinopyroxene oikocryst-bearing gabbro, and gabbronorite (Fig. F10).

The primary science results were obtained from the massive blocks drilled at re-entry Holes U1415J and U1415P located ~110 m apart and the single-bit Hole U1415I located ~10 m south of Hole U1415J (Figs. F8, F9). In the remainder of this section, only the results from the massive units from these holes are considered further (i.e., excludes the surficial rubble units).

Olivine gabbro is the dominant rock type recovered at Site U1415, occurring in both the layered series in Holes U1415I, U1415J, and U1415P and the Troctolite Series in Hole U1415P (Figs. F11, F12) (see below). Olivine gabbro is dominantly medium grained with equigranular–granular to granular–poikilitic textures and consists of olivine (5%–30%), plagioclase (45%–70%), and clinopyroxene (5%–45%), with trace amounts of orthopyroxene and oxides. Olivine is subhedral to euhedral with an equant to amoeboid skeletal habit, tabular plagioclase is euhedral to subhedral, and clinopyroxene is anhedral and dominantly subequant and often poikilitic. In Hole U1415P, multitextured olivine gabbros display complex variation in modal proportions of minerals, grain size, and mineral habit.

An unexpected observation is the prevalence of orthopyroxene-bearing lithologies, predominately orthopyroxene-bearing olivine gabbros and minor olivine-bearing gabbronorite (Figs. F11, F12). These medium- to coarse-grained, equigranular, granular rocks consist of olivine (10%–20%), plagioclase (50%–70%), clinopyroxene (15%–30%), orthopyroxene (1%–4%), and trace amounts of oxide (Cr-spinel). Olivine is subhedral to anhedral with an equant to amoeboid to skeletal habit, plagioclase is euhedral to subhedral with a tabular habit, clinopyroxene is anhedral forming interstitial grains, and orthopyroxene is euhhedral to anhedral forming subequant to equant prismatic or interstitial grains (Fig. F13). In some occurrences, orthopyroxene appears as an early phase, in which it is locally intergrown with olivine, followed by late-stage clinopyroxene (Fig. F14).

Troctolite is the second most abundant lithology, forming decimeter-scale intervals within the layered series and the dominant lithology in the Troctolitic Series (see below). In Hole U1415J, troctolites consist of olivine (20%–80%), plagioclase (20%–5%), clinopyroxene (<1%–10%), and trace amounts of oxide (possibly Cr-spinel). In Hole U1415P, troctolites have a more consistent modal mineralogy, with olivine (35%), plagioclase (65%), and trace amounts of clinopyroxene (<1%) and oxide (<1%, Cr-spinel). In both holes, olivine is euhedral to subhedral with a subequant to amoeboid to skeletal habit, plagioclase is euhedral to anhedral with a tabular habit, and clinopyroxene is anhedral with an interstitial habit. Troctolite has an equigranular and granular texture (Fig. F15).

Clinopyroxene oikocryst-bearing troctolite forms decimeter-scale intervals within the layered series in Holes U1415I and U1415J. The troctolites are medium grained, seriate poikilitic-granular rocks with a strong grain size contrast between minerals in the troctolitic matrix and chadacrysts in large clinopyroxene oikocrysts (Fig. F16A). The troctolites consist of olivine (10%–42%), plagioclase (45%–70%), and clinopyroxene oikocrysts (3%–35%), with trace amounts of oxide (possibly Cr-spinel) and orthopyroxene. Olivine is fine grained, subhedral to anhedral with an elongated, irregular amoeboid habit. Plagioclase is fine grained, subhedral to euhedral with a tabular habit. Large clinopyroxene oikocrysts (up to 15 mm across) are anhedral and poikilitic, with a distinctive population of irregularly shaped plagioclase chadacrysts (Fig. F17). The plagioclase chadacrysts are oriented in a random manner within the oikocrysts, in contrast to the surrounding foliated plagioclase fabric. Many of the plagioclase chadacrysts are deformed. Olivine is absent as a chadacryst phase.

Gabbros primarly occur in centimeter-scale intervals in the layered series in Holes U1415I and U1415J (Fig. F11) where they are fine to medium grained, and have an equigranular/granular texture. The primary mineralogy of the gabbros is dominated by plagioclase (55%–75%) and clinopyroxene (25%–45%) with trace amounts of olivine, orthopyroxene, and oxide. Plagioclase is euhedral to subhedral with a tabular habit, whereas clinopyroxene is anhedral with a subequant habit.

The rock types recovered from Holes U1415I, U1415J, and U1415P are grouped into lithologic units based on rock type and structural constraints (Figs. F11, F12); these units in turn may be simplified into three lithologic series. The first lithologic series includes the Layered Series in Holes U1415I and U1415J, which is composed of a ~30 m thick sequence of primitive olivine gabbronorites, olivine-orthopyroxene gabbros, and clinopyroxene oikocryst-bearing troctolites with moderate to strong magmatic foliation. The Layered Series is host to spectacular modal, and grain size layering. The boundaries between the layers are defined by changes in mineral modes and/or grain size and are primarily sutured such that individual mineral grains span the boundary (Fig. F18).

The second lithologic series is the Multitextured Layered Gabbro Series in Hole U1415P composed of a ~50 m thick sequence of olivine gabbros and orthopyroxene-bearing olivine gabbros with considerable banding, skeletal olivine textures, and occasional noritic bands (Fig. F12). This heterogeneous series reveals an extraordinary variety of textures that include variations in grain size, crystal morphology, and mineral fabrics and consists of a sequence of interfingered intervals that range from homogeneously textured primitive orthopyroxene-bearing olivine gabbros to dramatically banded or layered olivine gabbros. Banding can be divided into two broad categories: more common, steeply dipping, asymmetric and sometimes diffuse leucocratic banding and less common, more regular, gently dipping grain size and modal layering

The Troctolitic Series in Holes U1415J and U1415P comprise the third lithologic series (Figs. F11, F12). Melanocratic to leucocratic troctolites in this series are lithologically similar, commonly have a weak olivine-plagioclase foliation, and contain little modal layering (Fig. F15).

Basaltic and hypabyssal rocks

Basaltic and doleritic lithologies were largely recovered from the surficial rubble zone. The relationship between these rocks and the gabbroic units was only documented in the Troctolite Series in Hole U1415J where minor dikes are associated with zones of brittle deformation. The primary lithologies in Hole U1415N are moderately to highly olivine-phyric basalt and dolerite. Compositions of these basalts are at the primitive, depleted end of the EPR trend. Basalts found within the cataclastic intervals are discussed below.

Bulk chemistry

The gabbroic rocks from the layered, mutlitextured, and troctolite series are all primitive, with MgO contents ranging from ~7 to 31 wt% and Fe2O3T contents ranging from ~2 to 9 wt%. The olivine gabbros and gabbros have high Mg# (79–87) and Ca# (~77–92), high Ni contents (130–570 ppm), low TiO2 contents (0.1–0.3 wt%), and incompatible lithophile element contents (e.g., Y < 11 ppm). Five samples are distinguished by a significantly higher Cr content (~1500–2500 ppm) compared to neighboring gabbros (~36–825 ppm). Clinopyroxene oikocryst-bearing troctolites and gabbros from the Layered Series are similar in composition to the olivine gabbros. Orthopyroxene-bearing olivine gabbros have primitive compositions similar to neighboring gabbros and olivine gabbros and are characterized by high Mg# (80–85) and Ca# (77–87), high Ni (174–460 ppm) and Cr (150–2560 ppm) contents, low TiO2 contents (0.1–0.3 wt%), and trace element contents (e.g., Y < 6 ppm).

Troctolites overlap in composition with the gabbros but have, on average, more refractory compositions with high Mg# (81–89) and Ca# (79–98), high Ni (260–1500 ppm) and Cr (365–1100 ppm) contents, low TiO2 contents (<0.1 wt%), and incompatible lithophile element contents (e.g., Y < 3 ppm). The most refractory troctolites sampled in Hole U1415P have compositions overlapping the field of impregnated mantle peridotites. However, these samples are low in Ni relative to their high Mg#, indicating formation by a dominantly cumulate process (see below).

The gabbroic rocks at Site U1415 are far more primitive than the shallow-level gabbros at the Hess Deep Rift and are similar in bulk composition to gabbros from the shallow gabbros from Pito Deep where fast-spreading EPR crust is exposed (Perk et al., 2007). These primitive lithologies fall within the range of primitive oceanic gabbros from fast-spreading crust, as illustrated in a Mg# versus Cr content plot (Fig. F19). The Cr contents of these rocks likely reflect the crystallization of chrome spinel from a Cr-rich primitive parent magma.

Discussion of significant new observations

Layering

A layered lower crust is one of the key and nearly ubiquitous features of all models of the fast-spreading lower crust. Despite this, observations of modal and/or grain size layering in fast-spreading oceanic gabbros, indeed gabbros formed at all spreading rates, is rare and restricted to small intervals (e.g., Pito Deep [Perk et al., 2007], ODP Hole 735B [Dick, Natland, Miller, et al., 1999], IODP Hole U1309 [Blackman, Ildefonse, John, Ohara, Miller, MacLeod, and the Expedition 304/305 Scientists, 2006]). This lack of layering is in contrast to ophiolites where layering is nearly ubiquitous in plutonic rocks. Recovery of a wide range of layering observed in the Site U1415 gabbroic rocks is therefore of great significance, both for confirmation of its presence in fast-spreading crust and for its variety.

Two types of layering are seen at Site U1415: (1) simple centimeter- to decimeter-scale layering defined by differences in modal mineralogy and, more rarely, by grain size and (2) asymmetric and sometimes diffuse leucocratic banding (Fig. F20). Regular layering is largely defined by variation in modal olivine, plagioclase, and to a lesser extent pyroxene; the layers have sharp planar boundaries on a <1 cm scale (Fig. F20A, F20B). In Hole U1415J, some of the boundaries show grain size variation caused by the appearance of large (2–5 mm) pyroxene crystals and others by increases in olivine grain size from 1–2 to >5 mm. This type of layering is particularly well developed in intervals of olivine gabbro and gabbro in the Layered Series in Hole U1415I and the oikocryst-bearing Layered Gabbro Series in Hole U1415J. Weak to strong magmatic foliation is common in intervals with simple layering where its orientation is generally parallel to the layers.

Diffuse layering or banding, restricted to the olivine gabbro and gabbro intervals in the Multitextured Layered Gabbro Series in Hole U1415P, is defined by modal, grain size, and shape variations in plagioclase, olivine, orthopyroxene, and clinopyroxene and mainly manifests as variably distinct leucocratic and melanocratic bands (Fig. F20C). Magmatic foliation areas are poorly developed to absent in the Multitextured Layered Gabbro Series.

Layer boundaries are mainly sutured (Fig. F18) with minor gradiation, implying that they developed at hypersolidus conditions. The simple layering observed in Holes U1415I and U1415J is reminiscent of that found in basic layered intrusions (Workshop Participants, 1986). The diffuse banding in Hole U1415P is different, with textures and modal layering suggestive of magma mixing.

Orthopyroxene in primitive gabbroic rocks

The common occurrence of orthopyroxene in the primitive gabbroic rocks at Site U1415 was not expected. Orthopyroxene is a minor (<5%) and, in some centimeter-scale domains, a major (up to 25%) phase in olivine gabbro, gabbro, troctolite, and gabbronorite (Figs. F13, F14). These orthopyroxene-bearing rocks have Mg# ranging from ~0.80 to 0.85. Although gabbronorites are common and expected in evolved (average Mg# = ~0.58) shallow-level gabbros at the Hess Deep Rift (Gillis, Mével, Allan, et al., 1996; Natland and Dick, 1996; Coogan et al., 2002a), orthopyroxene has only once been observed in fast-spreading primitive gabbros (Coogan et al., 2002a).

The presence of orthopyroxene as a cumulate phase at Site U1415 is significant, as it is considered to be one of the last major minerals to crystallize from mid-ocean-ridge basalt (MORB)-like tholeiitic liquids (e.g., Stolper and Walker, 1980; Grove and Bryan, 1983; Grove et al., 1992; Feig et al., 2006). One model suggests that melts crossing the Mohorovicic seismic discontinuity have variable compositions compared to those used in phase equilibria experiments and that some may be in major but not trace element equilibrium with shallow depleted mantle (Coogan et al., 2002). This model led Coogan et al. (2002) to propose that some of the melt extracted from the mantle is fully aggregated within the mantle but reacts with the shallow mantle during melt extraction. Alternatively, phase relationships derived from experimental studies cited above may be different from those observed at Site U1415 because of differing conditions, emphasizing the importance of postcruise research to address this.

Olivine textures and their significance in characterizing strain

Intrinsic to layering observed at Site U1415 is the development of dendritic, hopper, and skeletal olivine textures in the cumulate rocks (Fig. F16B). These types of textures have been shown to be indicative of magmatic undercooling (O’Driscoll, 2007; Donaldson et al., 1982). Importantly, the presence of these olivine textures shows that deformation subsequent to crystallization was minimal, otherwise these delicate structures would be destroyed.

Clinopyroxene oikocrysts

The textures of clinopyroxene observed in gabbroic rocks at Site U1415 are remarkable for their complexity and diversity compared to previously sampled ocean crust. Textural habits range from equant granular clinopyroxene to centimeter-scale oikocrysts of poikilitic clinopyroxene. One particular clinopyroxene texture is so distinctive that, for the purposes of description, its host rock was named “clinopyroxene oikocryst-bearing troctolite.” Large poikilitic clinopyroxene oikocrysts commonly occur in a range of gabbroic lithologies recovered from elsewhere in the oceanic crust (e.g., Blackman, Ildefonse, John, Ohara, Miller, MacLeod, and the Expedition 305/305 Scientists, 2006; Cannat et al., 1995; Dick et al., 1999; Gillis, Mével, Allan, et al., 1993; Melson and Thompson, 1970); the clinopyroxene oikocrysts in the clinopyroxene oikocryst-bearing troctolites at Site U415 (Fig. F16A) are different. Here the clinopyroxene forms large, centimeter-scale, isolated oikocrysts contrasting strongly with a finer grained (millimeter scale) foliated troctolitic matrix. Plagioclase chadacrysts may be undeformed, bent with deformation twins and subgrains, or annealed; they may also be resorbed. No olivine chadacrysts were observed. These varying textures record a potentially complex hypersolidus deformation history of the crystal mush. We speculate that the origin of this distinctive clinopyroxene texture is related to melt migration and reaction within the lower oceanic crust. Similar lithologies have been well documented from particular horizons in tholeiitic layered intrusions in continental settings (Olmsted, 1979; Mathison, 1987), providing evidence for a commonality between some magmatic processes occurring at both layered intrusions and mid-ocean ridges, as originally proposed by Melson and Thompson (1970).

Hydrothermal alteration of fast-spreading East Pacific Rise lower crust

The metamorphic mineral assemblages in the rocks recovered at Site U1415 record the cooling of primitive gabbroic lithologies from magmatic (>1000°C) conditions at the EPR to zeolite (<200°C) facies conditions associated with Cocos-Nazca rifting and exposure onto the seafloor. The intensity of alteration varies with igneous lithology, in particular the modal abundance of olivine, grain size, and proximity to zones of brittle fracturing and cataclasis. The metamorphic assemblages in each lithologic interval record superimposed alteration conditions, the range of which records at what stage in the cooling of the section samples encountered fluids and deformation.

Pervasive alteration

The freshest lithologies at Site U1415 are found in the Layered Series of Holes U1415I and U1415J and Multitextured Layered Olivine Gabbro Series in Hole U1415P. The average alteration intensity in these units is ~40% (<10%–90%). Olivine is the most altered primary mineral (average = 55%–65%), followed by clinopyroxene and orthopyroxene (average = 30%) and plagioclase (average = 10%–20%). The Troctolite Series in Holes U1415J and U1415P are more pervasively altered than the gabbroic series, with Hole U1415J troctolites being more altered (~80%) than in Hole U1415P (~65%). This likely reflects the presence of brittle cataclastic zones in Hole U1415J (see below). Similar to the gabbroic series, olivine is the most altered primary mineral (65%–75%), followed by plagioclase (~60%) and clinopyroxene (35%–40%). In summary, away from zones of brittle deformation, alteration intensity is largely controlled by the bulk composition of the rocks such that the most olivine rich lithologies display the most alteration.

The secondary mineral assemblages replacing primary minerals display some differences between the gabbroic lithologies in the Layered Series, Multitextured Layered Olivine Gabbro Series, and Troctolitic Series (Fig. F21). These differences are primarily related to the relative abundance of the secondary phases replacing olivine and plagioclase. In all of the Layered Series, serpentine is slightly more abundant than talc, chlorite, and clay minerals. All assemblages have lesser amounts of amphibole, clay minerals, minor secondary oxides, and sulfides. In contrast to the Layered Series, olivine replacement in the Troctolitic Series is dominated by serpentine with lesser and varying proportions of chlorite and clay minerals and lesser talc, amphibole, minor secondary oxides, and sulfides. Similar to olivine, the relative abundance of the secondary minerals replacing plagioclase varies with rock type. The abundance of prehnite and chlorite are approximately equal in all of the gabbroic units, whereas prehnite is the most abundant secondary phase in the troctolites. In both rock types, secondary plagioclase, amphibole, and/or clay minerals also replace plagioclase. Secondary mineral assemblages replacing the other primary phases are not lithologically controlled. In all lithologies, clinopyroxene and orthopyroxene are largely replaced by amphibole and lesser chlorite.

A characteristic feature of altered olivine-rich gabbroic rocks is a concentric zonal aggregate of tremolite and chlorite that develops as a reaction product between olivine and plagioclase in oceanic gabbroic rocks (e.g., Nozaka and Fryer, 2011). Such corona textures are variably developed in all of the olivine-bearing gabbroic lithologies recovered at Site U1415 but the completion of the corona-forming action (i.e., absence of relict olivine inside a corona) appears most frequently in the Troctolitic Series in Hole U1415J.

Brittle fracturing and vein formation

Alteration veins represent a ubiquitous, although volumetrically insignificant, component of the rock types recovered at Site U1415. They reflect the later stages of cracking, fluid circulation, and fluid-rock reaction experienced by the gabbroic lithologies. Vein density is low except in the zones of extensive cataclasis found in Hole U1415J. The orientation of alteration veins is globally random; they form a network with no preferred orientation that is consistent with hydraulic fracturing.

A variety of vein mineralogies are observed at Site U1415, including amphibole, epidote, chlorite, serpentine, prehnite, carbonates, zeolite, and clay minerals. A systematic hierarchy in crosscutting relationships between veins with different fillings is hard to establish. Where present, amphibole and epidote veins are early, and mutually crosscutting relationships are observed for chlorite and prehnite. Zeolite veins are always late, crosscutting the other assemblages.

Discussion

Pervasive alteration in the gabbroic lithologies spanned amphibolite to subgreenschist facies conditions (from >700° to <200°C) (Fig. F22). Clear evidence for early high-temperature amphibolite facies is rare and is manifested by the rare occurrence of brown amphibole and green spinel associated with olivine replacement in Hole U1415J. Some of the secondary amphibole associated with pyroxene and plagioclase replacement may also record amphibolite facies, by analogy with gabbroic rocks from elsewhere at Hess Deep, but must be confirmed by postcruise analyses. Incipient alteration in the majority of the recovered core occurred at lower amphibolite (450°–650°C) facies based on corona textures and olivine alteration assemblages (e.g., Frost et al., 2008; Nozaka and Fryer, 2011). The predominant alteration occurred at greenschist (<400°C) to subgreenschist (<200°C) facies, as evidenced by serpentine, chlorite, and prehnite assemblages and zeolite and clay mineral assemblages, respectively.

Pervasive alteration in the Site U1415 primitive gabbroic rocks differs from the more evolved, shallow-level gabbros found at the crest of the intrarift ridge and northern escarpment at Hess Deep. On average, these shallow-level gabbros are fresher, particularly in comparison with the troctolites, and the dominant alteration is amphibole after pyroxene and minor secondary plagioclase and amphibole and chlorite after igneous plagioclase (Früh-Green et al., 1996; Gillis, 1995; Kirchner and Gillis, 2012). As olivine is absent or of low modal abundance in the shallow gabbros, abundant serpentine and talc would not be expected. Moreover, the lack of prehnite and chlorite replacing plagioclase may also be related to olivine abundance, as prehnite and chlorite may be by-products of the serpentinization process described above. Another feature that is ubiquitous, though with variable abundance, in the shallow gabbros and rare at Site U1415 is amphibole filling microfractures and lining grain boundaries at amphibolite facies (>700°C). An important question to address is whether incipient fracturing at amphibolite facies is rare in fast-spreading primitive gabbros, or if mineral assemblages differ because of their more evolved bulk compositions (e.g., McCollom and Shock, 1998), or if the observations at Site U1415 simply reflect local scale heterogeneity caused by, for example, proximity to a zone of focused fluid flow (Coogan et al., 2006).

Comparison with the vein types found at Site 894 provides some constraints on the evolution of fracture-controlled fluid flow. Veins in the Site 894 cores are similar to those at Site U1415, from oldest to youngest, with amphibole, chlorite ± prehnite ± epidote, chlorite–clay minerals, and zeolite veins (Früh-Green et al., 1996; Manning and MacLeod, 1996). The chlorite and zeolite vein types have moderate to steep dips to the south that show a strong preferred east–west orientation when reoriented based on magnetic and FMS results (Manning and MacLeod, 1996). Because the EPR strikes north–south at this latitude, the veins’ east–west orientations suggest formation by a mechanism unrelated to fracturing in an EPR hydrothermal system. Although the dominant fracture sets found in the Site 894 core were not observed at Site U1415, similarity in the greenschist to subgreenschist facies vein assemblages suggests a common evolution.

Evolution of cataclastic deformation, magmatism, and hydrothermal alteration

Another significant result of Expedition 345 is the recovery of blocks hosting a record of cataclasis whose development was synchronous with basaltic dike intrusion. The fault zone occurs within in the Troctolite Series of Hole U1415J, where the recovered core documents complex interactions between cataclastic faulting, fluid flow and alteration, and magmatism. Cataclasis is largely localized at two depth intervals in the Troctolite Series, where brittle fabrics range from dense-anastomosing fractures to well-developed breccias and cataclasite. Elsewhere in Holes U1415J and U1415P, only very thin (less than centimeter scale) zones of fracturing and cataclasis were recovered.

The zone of brittle deformation in Hole U1415J exhibits heterogeneous grain sizes and degree of alteration and reflects the variable intensity of cataclasis over a centimeter scale. Cataclastic fabrics are characterized by grain size reduction through microcracking and rotation of the primary igneous and metamorphic minerals. This microstructure is commonly cut by prehnite and minor chlorite veins that are in turn cut by another period of cataclastic deformation. Locally, prehnite and chlorite veins are deformed and crosscut by later prehnite and chlorite veins. These relations indicate a complex succession of vein formation and brittle deformation.

Intervals of core host evidence of localized dike intrusion synchronous with the development of fault rocks in Hole U1415J. This is based on the combined evidence from textures of several pieces of fractured dolerite, two of which have very fine grained margins adjacent to altered locally cataclasitic gabbro. As an example, microstructural observations suggest that one dolerite was likely emplaced into cohesive gabbro cataclasite hosting variable intensity prehnite/chlorite alteration. Boundaries with the cataclastite are sharp, with a distinctive increase in grain size away from the contact. Similar to elsewhere in the zones of fracture and cataclasis, the dolerite, contact zone, and cataclastic gabbro are cut by epidote and chlorite veins that were deformed by later brittle deformation. These relationships suggest the following sequence of events: (1) localized fracture and cataclasis and fluid flow associated with faulting and low-temperature (<400°C) vein formation, (2) vein intrusion, (3) further cataclasis, (4) dolerite dike emplacement, and (5) another phase of brittle fracture.

A critical unresolved question is the timing of faulting and dike injection, as this would determine whether these processes are associated with the EPR or Cocos-Nazca Ridge. No results from the Expedition 345 can readily resolve this issue. However, comparison with the Hole 894 core may provide some constraints, as localized zones of cataclasite and vein networks with similar textures and alteration mineral assemblages to those found in Holes U1415J and U1415P were recovered. Magnetic and FMS data from Hole 894G indicate that veins and fractures show a strong preferred east–west strike and steep south-facing dips (MacLeod et al., 1996b). As discussed above, similarity in the greenschist to subgreenschist facies vein assemblages suggests that brittle deformation and alteration may be associated with Coco-Nazca rifting.

Paleomagnetism

Paleomagnetic analyses focused on determining the structural continuity of units sampled in Holes U1415J and U1415P. Variations in inclination of stable, high unblocking temperature magnetic components indicate that at least two blocks had been sampled in each hole, providing evidence of relative displacements of individual, internally coherent units (Fig. F23). Thermal demagnetization experiments also identified more complex remanences in some samples from Hole U1415J, indicating that magnetizations were acquired during at least two geomagnetic polarity chrons (Fig. F24). Postcruise paleomagnetic research will focus on understanding the origin of these multipolarity remanences.