The seafloor of the southernmost Cocos plate is characterized by morphological, structural, and thermal segmentation resulting from a >20-m.y. history of plume-ridge interactions and ridge jumps of the CNS. A "plate suture" exists offshore Costa Rica where lithosphere formed at the fast-spreading (~130 mm/yr full spreading rate) (Wilson, 1996) East Pacific Rise (EPR) is juxtaposed against lithosphere generated at the intermediate-spreading (65 mm/yr) (Barckhausen et al., 2001) CNS (Figs. F1, F2). Crust at the location of Legs 170 and 205 is clearly identified from the magnetic anomalies as having been generated at the EPR and as being ~24 Ma old. The initiation of CNS spreading may have occurred when a preexisting transform fault on the Farallon plate passed over the hotspot at ~24 Ma (Barckhausen et al., 2004). Crust generated at the EPR is characterized by a smooth morphology, whereas the CNS crust is morphologically irregular and has high-amplitude magnetic anomalies. The transition at this "rough/smooth" boundary (Hey, 1977) occurs in an area of abundant seamounts (~40% coverage) (von Huene et al., 1995). The spreading history of the CNS is characterized by multiple ridge jumps and migrations associated with differences in local plate motions and hotspot-induced ridge reset. The long-lived (~90 m.y.) (Hauff et al., 1997) Galápagos hotspot has continuously overprinted regional volcanic products, as evidenced in the geochemistry of the Galápagos Islands (White et al., 1993), abundant seamounts (e.g., Fisher Seamount and the Costa Rica seamount province) (Harpp and White, 2001; Werner et al., 2003), and forearc basaltic complexes in Central America (e.g., Nicoya, Osa, and Quepos) (Hauff et al., 2000). The prominent Cocos and Carnegie aseismic ridges, located on the Cocos and Nazca plates, respectively, are a consequence of the CNS location above the hotspot for at least the last 20 m.y. (Hey, 1977; Meschede and Barckhausen, 2001).
Segmentation of the Cocos plate is reflected in the structure of the overriding Caribbean plate (von Huene et al., 2000), upon which the Central American volcanic arc (CAVA) is constructed. Considerable control of Quaternary convergent margin tectonics and variable arc lava geochemistry are attributable to the subduction of different crustal types of the Cocos plate that result partly from plume-ridge interaction (Patino et al., 2000; Ranero and von Huene, 2000; von Huene et al., 2000). For example, a trench-normal fracture associated with a CNS ridge propagator (created after the CNS-1 to CNS-2, ~19.5 Ma ridge jump) (Fig. F2) continues into already subducted lithosphere (Barckhausen et al., 2001), and its landward extension coincides with an abrupt offset in the depth-to-slab beneath the CAVA called the Quesada sharp contortion (QSC) (Protti et al., 1995) (see Fig. F2 for location). The QSC also coincides with an arc gap between Arenal and Platanar-Porvenir-Poás stratovolcanoes in Costa Rica (Protti et al., 1995) and an abrupt change in the slab (sediment + oceanic basement) signal in the geochemistry of arc lavas (Carr et al., 2003, and references therein; Feigenson et al., 2004).
The Cocos plate was drilled across the MAT offshore Costa Rica during Legs 170 and 205 (Kimura, Silver, Blum, et al., 1997; Morris, Villinger, Klaus, et al., 2003). Patterns of magnetic anomalies indicate that the basement was formed at the EPR at ~24 Ma (Barckhausen et al., 2001), and precruise seismic profiles (Shipley et al., 1992) suggested a depth-to-basement of ~300 m below the seafloor/sediment interface (Fig. F3); drilling proved it to be at 378–422 m. Drilling during Leg 170 bottomed in an igneous unit after coring ~30 m at Site 1039 and 8 m at Site 1040. This igneous unit was interpreted as multiple apophyses of sill-like magma injections based on petrographic and operational data (Kimura, Silver, Blum, et al., 1997). Subsequent drilling at nearby Site 1253 (~1 km from Site 1039) during Leg 205 penetrated this unit (Subunit 4A), cored an additional 30 m of sediment, and encountered a second igneous unit (Subunit 4B). Drilling operations ended at this site after coring ~150 m of Subunit 4B without reaching its lower boundary. The recovery of igneous rocks at Site 1253 in relation to lithostratigraphic units is presented in Figure F4. Similar to Subunit 4A, Subunit 4B appears to represent a sill intrusion (see below). At Sites 1039 and 1253, the intrusion age of Subunit 4A is post-15.6–18.2 Ma derived from biostratigraphic ages of the intruded sediments (Muza, 2000). Dating is in progress for Subunit 4B.
Shipboard and postcruise thin section analyses performed on igneous rocks recovered from Legs 170 and 205 indicate that they are petrographically similar to oceanic igneous rocks cored from nearby Deep Sea Drilling Project (DSDP) Hole 504B on the Nazca plate and ODP Site 1256 on the Cocos plate (Fig. F1). Subunits 4A and 4B are microcrystalline to fine-grained and occasionally medium-grained plagioclase-clinopyroxene gabbro with plagioclase aggregates (glomerocrysts) or, more rarely, plagioclase with pyroxene aggregates in a nearly holocrystalline (microcrystalline to fine-grained, and, rarely, medium-grained) groundmass (Kimura, Silver, Blum et al., 1997; Morris, Villinger, Klaus et al., 2003). Minor primary phenocryst phases include highly altered olivine, rare orthopyroxene, ilmenite, and magnetite. Secondary phases identified by X-ray diffraction within the igneous section include zeolites, chlorite, vein-filling calcite, smectite, and products of glass alteration tentatively identified as saponite and celadonite suggesting low-temperature alteration (Kimura, Silver, Blum et al., 1997; Morris, Villinger, Klaus et al., 2003).
The crystalline texture suggests that these rocks do not necessarily preserve original melt compositions, and plagioclase growth zoning indicates magma differentiation during crystallization of phenocrysts. Phenocryst assemblages with contrasting morphologies are contained within groundmass domains of different grain-sizes, commonly separated by apparent magmatic contacts (Morris, Villinger, Klaus et al., 2003), supporting the interpretation that these subunits are multiple magma injections into partially crystallized microcrystalline gabbro.
Subunit 4A was further divided into two subunits and Subunit 4B into seven subunits based on the distribution of voids, veins, grain size variation with depth and the proportions of plagioclase to pyroxene (Morris, Villinger, Klaus, et al., 2003). Discrete alteration is greatest at the tops of the subunits and is generally higher in the lowermost cores at 1–5 vol%; some horizons in lower Subunit 4B reach nearly 50 vol% alteration. At 513 meters below the seafloor (mbsf) within Subunit 4B, a thin basaltic (cryptocrystalline) interval was recovered. Below this depth, Subunit 4B contains a higher abundance of glass and magmatic contacts, is generally more altered, and has fractures of larger size. Shipboard scientists debated the possibility that the igneous section below this horizon could be extrusive EPR oceanic crust with the overlying material resulting from shallow sill emplacement. Geochemical evidence presented here, however, argues that all igneous material cored during ODP Leg 205 results from magmatic overprinting of the EPR basement that was generated from an enriched source. Hydrothermal alteration and implications of fluid flow within Subunit 4B are discussed elsewhere (Dreyer et al., 2005).