Abstract

At the present, the geochemical influence of the Galápagos hotspot (offshore South America) can be seen only along the Galápagos spreading center, north of the hotspot. It is possible, however, that Galápagos plume material also reached the East Pacific Rise in the past. Detecting such influence would be of particular importance for the interpretation of geochemical data from oceanic crust at Ocean Drilling Program (ODP) Site 1256, which formed ∼15 m.y. ago at the East Pacific Rise during a Miocene period of superfast spreading, and is considered to be a reference site for oceanic crust produced at fast-spreading ridges. Here we present geochemical data from Miocene basaltic crust (23–7 Ma) drilled at several Deep Sea Drilling Project (DSDP), ODP, and Integrated Ocean Drilling Program (IODP) sites that formed along the East Pacific Rise between 3°S and 7°N. Lavas formed between ca. 22.5 and ca. 11 Ma show enriched, Galápagos plume–like Pb and Nd isotope ratios (with a peak in enrichment between ≥18 and 12 Ma) compared to lavas created shortly before or after this time interval. Despite their enriched isotope composition, these samples generally show depletion in more-incompatible, relative to less-incompatible, trace elements. Derivation from an enriched Galápagos plume source that had experienced recent melt extraction before it melted further beneath the East Pacific Rise can explain the combined incompatible-trace-element depletion and isotopic enrichment of the 22.5–11 Ma lavas. The influence of plume material correlates with the interval of superfast spreading along the equatorial East Pacific Rise, suggesting a causal relationship. Enhanced ridge-plume interaction (“ridge suction”) due to superfast spreading could have facilitated the flow of Galápagos plume material to the ridge. On the other hand, the arrival of Galápagos-type signatures took place immediately after formation of the Galápagos spreading center, which could have provided a pathway for hot plume material to spread into the main ridge network.

INTRODUCTION

Since the pioneering work of Schilling (1973), interaction of hotspots or mantle plumes with mid-ocean ridges has been documented worldwide (e.g., Schilling et al., 1982; Hanan and Schilling, 1989; Schilling et al., 1994; Jones et al., 2002; Ito et al., 2003). However, the dynamics of mantle flow and melting remain a matter of debate, in particular for off-axis plumes, where interaction with nearby ridges is less obvious (e.g., Small, 1995). One of the best examples of plume-ridge interaction is between the Galápagos hotspot and the Galápagos (or Cocos-Nazca) spreading center (GSC) in the eastern equatorial Pacific Ocean (Fig. 1). The Gálapagos hotspot lies ∼200 km south of the GSC, which was initiated by the break-up of the Farallon plate, possibly caused when a preexisting fracture zone passed over the Gálapagos hotspot ∼23 m.y. ago (Handschumacher, 1976). The relative position of the GSC to the hotspot changed several times, from being hotspot-centered to slightly south or north of the hotspot until it eventually shifted to north of the hotspot between ca. 11 Ma and 8.2 Ma (Werner et al., 2003; Wilson and Hey, 1995). At present, the GSC forms a triple junction with the East Pacific Rise (EPR) ∼1100 km west of the hotspot.

Systematic variations in basalt chemistry along the GSC near the Galápagos Archipelago (e.g., an increase in incompatible-element abundance and isotopic enrichment, expressed by more radiogenic [higher] Pb and less radiogenic [lower] Nd isotope ratios) are explained by transport of plume material from the hotspot to, and then along, the spreading center (Christie et al., 2005; Detrick et al., 2002; Kokfelt et al., 2005; Schilling et al., 2003). At present, the geochemical influence of the hotspot can only be detected as far west as 95.5°W (500 km from the hotspot) with major and trace elements, volatiles, and He isotopes (Detrick et al., 2002) or 99.5°W (1000 km from the hotspot) with Sr, Nd, Pb, and Hf isotopes (Schilling et al., 2003). It can be speculated, however, that Galápagos plume material might have reached the EPR in the past (e.g., Sadofsky et al., 2009), e.g., due to enhanced “plume push” or increased “ridge suction” at times of higher spreading rates. Whereas the current full spreading rate of the EPR in this region is ∼130 mm a−1 (Lonsdale, 1988), a period of superfast spreading of 180–210 mm a−1 between 11 and ≥18 Ma is inferred from magnetic anomalies (Wilson, 1996). In contrast, the maximum inferred (asymmetric) full spreading rate for the GSC was 62 mm a−1 (Meschede and Barkhausen, 2001), slowing to 20–40 mm a−1 at ca. 10.5–11 Ma.

The question of whether Galápagos plume material reached the EPR is of particular importance for the interpretation of data from oceanic crust at Ocean Drilling Program (ODP)/Integrated Ocean Drilling Program (IODP) Site 1256, which was formed ∼15.2 m.y. ago at the EPR during a period of superfast spreading near today’s triple junction with the GSC (Fig. 1) (Wilson, 1996). Hole 1256D, which was initiated during ODP Leg 206, and meanwhile re-occupied and deepened by three successive IODP expeditions, is considered to be a reference site for oceanic crust formed at fast-spreading ridges. A possible Galápagos influence could raise the question as to how geochemically representative the Site 1256 basalts actually are (Duggen et al., 2008; Park et al., 2008; Sadofsky et al., 2009). To address this question, and to evaluate if there is a connection between superfast spreading and influx of the Galápagos hotspot mantle into the EPR, we have analyzed basaltic crust samples, formed between 3°S and 7°N at the EPR between 7 and 23 Ma, from Deep Sea Drilling Project (DSDP) Sites 82, 83, and 495; ODP Site 844, 845, 846, 849, 851, and 1226; and ODP/IODP Site 1256.

RESULTS

We determined the trace element and Sr, Nd, and Pb (double-spike) isotope compositions for 14 basaltic whole-rock samples from the top of the igneous basement from several paleoceanographic DSDP/ODP sites, and for 13 fresh glass samples from the upper basaltic sequence at the deep crustal drill Site 1256 (see Fig. 1 for site locations, and Table DR1 in the GSA Data Repository1 for exact coordinates, inferred ages, and references). To facilitate comparison between studies, we also reanalyzed four whole-rock samples from this site that were previously analyzed with the conventional Pb isotope technique by Sadofsky et al. (2009). Details of the preparation and analytical procedures, and trace element and isotopic data tables, are given in Tables DR2–DR4.

All samples, with two exceptions, have (La/Ce)n, (La/Sm)n, and (La/Yb)n <1, overlapping and falling below the range for normal mid-oceanic ridge basalt (N-MORB) (Fig. 2A). A whole-rock sample from Site 845 and a glass sample from Site 1256 display slightly elevated highly over moderately incompatible element ratios [e.g., (La/Yb)n = 1.1–1.2] plotting within the Galápagos range. Ratios of intermediate to heavy rare earth elements (REE) such as (Tb/Lu)n also overlap with N-MORB or extend to lower values (Fig. 2B).

With the exception of the oldest samples from Sites 499/500 (ca. 22.5–23.0 Ma; data previously published by Geldmacher et al., 2008), all samples have more enriched Pb and Nd isotopic compositions than nearby (9°–10°N of the EPR) recent N-MORB, and fall within or very close to the field for the Galápagos Islands (Figs. 2C and 2D). It is important to note that the Nd and Pb isotopic composition of the whole-rock samples from Site 1256 overlap with the fresh glass analyses from this site, showing that these isotope systems have not been substantially affected by alteration in the whole-rock samples.

Systematic variations exist between sample ages and their incompatible-element and isotope ratios (Figs. 3A and 3B). Samples from Sites 499/500 (ca. 23 Ma; Geldmacher et al., 2008) have Pb and Nd isotopic and (Tb/Lu)n ratios similar to local N-MORB. Samples from Site 495 (located only 24 km from Sites 499/500 on ∼0.3 m.y. younger crust) and a sample from a small seamount (SO144–1; Werner et al., 2003) offshore of Nicaragua, formed on- or near-axis and thus having an age of ca. 22.5–23.0 Ma (based on the magnetic anomaly C6Cn of the underlying crust), have slightly more enriched isotopic compositions but also slightly more depleted (Tb/Lu)n, reflecting depletion of middle REEs relative to heavy REEs. Samples from Site 844 (17.3 Ma) have the most-enriched isotopic ratios but more-depleted (Tb/Lu)n ratios, while samples from Site 1256 (15.2 Ma) display slightly less-enriched isotopic ratios but the lowest (Tb/Lu)n ratios. With decreasing age from ca. 12 to ca. 7 Ma, samples from Sites 851, 849, 83, and 82 show systematically less-enriched isotopic compositions, but increasing (Tb/Lu)n and other incompatible-element ratios.

DISCUSSION AND CONCLUSIONS

The isotope and incompatible trace element data for the ca. 22.5 to ca. 11 Ma lavas formed at the EPR between 4°S and 7°N do not have typical MORB compositions. Instead, the Pb and Nd isotopic compositions fall within the Galápagos hotspot field, whereas the incompatible-element compositions are similar to, or even more depleted than, EPR N-MORB. Lavas formed between 11 and ca. 9 Ma display a transitional composition. Note that the observed geochemical anomaly also deviates from nearby E-MORB isotopic and trace element compositions (Fig. 2), and therefore cannot be explained by normal upper-mantle heterogeneity. If Galápagos plume material is responsible for the enriched isotopic composition of the 22.5–11 Ma EPR basalts, three possible processes need to be considered: (1) increased ridge suction, (2) increased plume push, or (3) change in ridge geometry.

As shown by Niu and Hékinian (2004), ridgeward mass flux, called “ridge suction force,” increases linearly with increasing spreading rate. Therefore, superfast spreading could have increased the suction of the regional upper mantle, and may have been strong enough to draw Galápagos plume material to the EPR between 3°S and 7°N.

Alternatively, the arrival of Galápagos plume material may be caused by an increased plume upwelling between 23 and 11 Ma, as deduced by the preserved Galápagos hotspot tracks on the Cocos and Nazca plates (Fig. 1): the volume of the combined Cocos and Carnegie Ridges between ∼11–15 m.y. ago is significantly greater than after 11 m.y. ago (see Werner et al., 2003), which was also the case for the combined Carnegie, Malpelo, and Coiba Ridges from ∼15 to >17 m.y. ago (ages from Hoernle et al. [2002] and O’Connor et al. [2007]). The older parts of these tracks, however, are no longer preserved on the seafloor, and therefore no information is available for the arrival time of plume material at the EPR.

The arrival of the isotopically enriched Galápagos-type mantle at the EPR at ca. 22.5 Ma coincides with the breakup of the Farallon plate ∼23 m.y. ago (e.g., Lonsdale, 2005; Meschede and Barkhausen, 2001). The newly formed GSC could have provided a passageway for mantle flow from the Galápagos hotspot to the EPR against the direction of asthenospheric flow and plate motion. When the GSC spreading changed from intermediate (∼60–70 mm a−1) to slow (∼20–40 mm a−1) at ca. 10.5–11 Ma (manifested by the rough-smooth boundary between crust formed at the EPR and the GSC since this time; Fig. 1), efficient along-axis flow of plume material was no longer promoted and the plume signal in EPR lavas disappeared (Fig. 3).

Presumably, a combination of these processes is most likely. The Farallon plate breakup serves as an upper bound for the beginning of superfast spreading along the equatorial EPR (Wilson, 1996). Superfast spreading, however, is well constrained between 18 and 11 Ma, correlating with a clear isotopic enrichment and incompatible-element depletion along the equatorial EPR (Fig. 3). Because today’s isotopic enrichment along the GSC can be traced for more than 1000 km (as far as 99.5°W, which is only ∼230 km away from the junction with the EPR; Fig. 1), it is likely that Galápagos plume material, spreading beneath the newly formed GSC, entered the EPR during times of greatly enhanced ridge production due to superfast spreading, thereby displacing the regular asthenospheric MORB source mantle beneath the ridge. This model is also consistent with the disappearance of the plume signal in EPR lavas at ca. 10 Ma, shortly after the end of superfast spreading at 11 Ma (Fig. 3), when GSC spreading also slowed down, hampering further along-axis flow of plume material. On the other hand, the apparent waning of the isotope signal after ca. 17 Ma (Fig. 3) could also reflect the increasing distance of a westward-migrating EPR from the hotspot. Past absolute distances between the EPR and the Galápagos hotspot, however, are difficult to constrain (Lonsdale, 2005), and the EPR might not have migrated much to the west at all (Rowley et al., 2011).

The apparent paradox of isotopic enrichment but incompatible-element depletion of the 22.5–11 Ma EPR lavas can be explained by previous depletion of the contaminating plume material through partial melting and melt extraction, either under ancient volcanoes produced by the Galápagos hotspot (Hoernle et al., 2000; Park et al., 2008) or during its flow along the GSC before reaching the EPR. The more-fertile, incompatible-element and isotopically enriched components (e.g., garnet pyroxenite/eclogite) of a heterogeneous plume (as has been proposed for the Galápagos; e.g., Harpp and White, 2001; Hoernle et al., 2000; White et al., 1993) would have gotten progressively exhausted through ongoing melt extraction in the deeper melting regions beneath the ridge system, leaving the more refractory and isotopically less-enriched plume matrix behind (Niu and Hékinian, 2004; Shorttle et al., 2010). By the time the residual material reached the EPR, it would have been strongly depleted in incompatible elements, and thus possess lower, more-incompatible over less-incompatible element ratios than lavas from the Galápagos Islands (Fig. 2A). In addition, the exhaustion of more-fertile lithologies containing garnet (which retains heavy REEs) would result in ratios of light to heavy REEs, such as Tb/Lu, to be similar to or even lower than the MORB source (Figs. 2B and 3B), as illustrated by the melting models shown in Figure 2B. The isotopic composition of the residual depleted plume matrix, however, would plot at the depleted end of the Galápagos field (Figs. 2C and 2D). The observed geochemical anomaly in the EPR lavas, and its temporal correlation with regional ridge tectonics, strongly imply an interrelationship between the formation of the GSC, arrival of Galápagos-type material at the EPR, and superfast spreading.

We thank S. Hauff for analytical support, S. Herrmann for updating the nannofossil ages (in Table DR1), and M. Portnyagin for fruitful discussions. The manuscript benefited from constructive comments by W. White, D. Geist, and one anonymous reviewer. We thank W. Collins for editorial handling. This research used samples provided by the Integrated Ocean Drilling Program (IODP). Funding for this study was provided by the Deutsche Forschungsgemeinschaft (DFG) IODP priority program grants HO1833/16-1 and HO1833/18-1.

1GSA Data Repository item 2013045, analytical details, paleomagnetic constraints, details of modeling, and MORB references; Table DR1 (site location, crustal ages and references); Table DR2 (whole-rock trace element and isotopic compositions); Table DR3 (glass trace element composition); and Table DR4 (glass isotope compositions), is available online at www.geosociety.org/pubs/ft2013.htm, or on request from editing@geosociety.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.