The Galápagos Archipelago is located ∼1000 km west of South America on the Nazca plate, which is moving eastward at 51 km/Ma (Argus et al., 2011). The east-west trending Galápagos Spreading Center (GSC), where the Nazca and Cocos plates are diverging, is between 150 and 300 km north of the archipelago (Figure 1A). The GSC has experienced multiple southward ridge jumps during the last 5 million years, the most recent at ∼1 Ma. Consequently, the GSC was closer to the Galápagos plume when the central and eastern islands were forming than it is today (Wilson and Hey, 1995; Mittelstaedt et al., 2012). Santa Cruz is one of several older islands in the central Galápagos, in addition to San Cristóbal, Santiago, Española, Floreana, Rabida, Pinzón, and Santa Fe (Figure 1A). Santiago, Santa Cruz’s neighbor to the northwest, has experienced limited historical activity, having erupted as recently as 1906 (Siebert et al., 2010), and San Cristóbal has ∼5 ka lavas along its northern coastline (Geist et al., 1986; Mahr et al., 2016).

The geochemistry of Galápagos lavas is complex; lavas ranging from depleted, MORB-like compositions to enriched alkali-olivine basalts have erupted across the archipelago (e.g., White and Hofmann, 1978; Geist et al., 1988; Graham et al., 1993; White et al., 1993; Kurz and Geist, 1999; Blichert-Toft and White, 2001:; Harpp and White, 2001). White and Hofmann (1978) first noted the distinctive, horseshoe-shaped spatial distribution of geochemical signatures in the Galápagos; along the western and southern periphery of the archipelago, volcanoes erupt lavas with geochemically enriched signatures, whereas the central and eastern volcanic centers produce material with depleted, MORB-like signatures (e.g., Geist et al., 1988; White et al., 1993; Harpp and White, 2001).

Initially, the enriched geochemical horseshoe pattern was attributed to plume-asthenosphere mixing (Geist et al., 1988) or thermal entrainment of the asthenosphere into a sheared plume (White et al., 1993; Blichert-Toft and White, 2001; Harpp and White, 2001). Researchers have also invoked a spatially heterogeneous plume, in which multiple mantle reservoirs with different evolutionary histories supply the plume (Hoernle et al., 2000; Gibson et al., 2012; Harpp et al., 2014b; Harpp and Weis, 2020; Gleeson et al., 2021). Harpp and White (2001) established that at least four isotopically distinct mantle endmembers are required to generate the geochemical variation observed across the archipelago: 1) plume material with primordial helium isotope signatures (PLUME; Graham et al., 1993), supplying the western shields; 2) ancient recycled material with fluid-mobile element-enriched signatures, concentrated along the southern archipelago near Floreana Island (FLO); 3) moderately depleted material with elevated ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb signatures, limited to the northern reaches of the archipelago and whose source remains enigmatic (WD); and 4) the depleted upper mantle (DUM). Harpp and Weis (2020) proposed that the more enriched signatures originate in the Large Low Shear Velocity Province (LLSVP) at the core-mantle boundary, incorporating ancient, recycled material to provide the enriched signatures, whereas the central and eastern lavas are supplied primarily by less enriched, lower Pacific mantle. They proposed that the compositional boundary between the W/SW enriched and E/NE depleted zones of the bilaterally asymmetric plume runs ∼NW-SE, cutting through the archipelago west of Santa Cruz (Figure 1A).

The archetype model for hotspot-generated ocean island evolution is based on Hawai’i (MacDonald and Katsura, 1964; Chen and Frey, 1983; Macdonald et al., 1983; Clague and Dalrymple, 1987; Clague and Dalrymple, 1988). Hawaiian volcanoes experience four lifecycle stages, explained by their proximity to the plume as the edifice is carried northwest by the Pacific plate: 1) the pre-shield stage, when the volcano is on the upstream periphery of the plume, generating low-degree melts with alkalic compositions; 2) the shield stage, dominated by high-volume, high-degree melts and tholeiitic lavas; 3) the post-shield phase, primarily alkalic, low degree melts at the downstream edge of the plume; and 4) rejuvenated volcanism, which occurs after ∼0.25–2.5 Ma of quiescence (e.g., Clague and Dalrymple, 1987, 1988).

Galápagos volcanoes do not conform to the Hawaiian evolutionary sequence (Geist et al., 2014a; Harpp and Geist, 2018; Harpp and Weis, 2020). For example, the central and eastern Galápagos volcanoes do not exhibit clearly defined post-shield or rejuvenation stages (e.g., Geist et al., 1986; Gibson et al., 2012; Harpp et al., 2014a). Furthermore, the eastern Galápagos volcanoes bear little morphological resemblance to the current western shields. The older islands of Santiago, Santa Cruz, San Cristóbal, and Española lack submarine rift zones and calderas characteristic of many western shields (and most Hawaiian volcanoes) and lack buried calderas (Cleary et al., 2020). Like Hawaiian volcanoes, however, the eastern volcanoes also experience extended volcanically active phases, erupting for >2 m.y. beyond when the island was located near the plume center (e.g., Geist et al., 1986; White et al., 1993; Mahr et al., 2016). Finally, petrographic modeling by Geist et al. (1998) indicates that magmatic cooling in the eastern islands, including Santa Cruz, is primarily controlled by clinopyroxene rather than plagioclase. They conclude that the bulk of the fractionation in eastern Galápagos island melts is taking place in the deep crust, at depths within the mantle in the range of 17–23 km. This depth interval is greater than that for the young western shields, whose magmatic plumbing systems are at shallower depths (∼10 km; Geist et al., 1998; 2014a) owing to their more robust magma supplies.

As a consequence of the striking differences between the eastern and western Galápagos volcanoes, Harpp and Geist (2018) proposed that the eastern islands are magmatically starved compared to the western shields, preventing development of thermochemically buffered magmatic systems capable of supporting long-lived magma chambers and caldera formation (Geist et al., 2014a). They attribute the lower magma flux to the proximity of the GSC between 1 and 5 Ma (e.g., Wilson and Hey, 1995; Mittelstaedt et al., 2012), which may have diverted plume material toward the ridge, away from the volcanoes developing at the time, which are currently the older, central and eastern Galápagos islands.

As the largest island in the center of the archipelago, Santa Cruz provides an opportunity to document how ocean islands mature as they are carried downstream from a mantle plume. Three previous field studies of Santa Cruz’s development (Bow, 1979; Schwartz, 2014 M.Sc. thesis; Schwartz et al., 2022; the original Masters’ thesis of Schwartz is available here: https://drive.google.com/file/d/1iMk9grGWRNORT7K2JxjsoXs4EY6IPKRU/view?usp=sharing), conclude that Santa Cruz has experienced two periods of volcanism and two periods of tectonic activity since the volcano migrated east from the hotspot center during the last >1 million years, as follows (Ar-Ar ages all from Schwartz et al., 2022):

1. Formation of the Platform Series, emplaced primarily between 1620 ± 15 and 1160 ± 35 ka, with only a trivial volume of activity extending beyond 1160 ka.

The Platform Series consists of lavas exposed primarily along the northeastern coast of Santa Cruz (Figure 1B), as well as Baltra and Seymour Islands, which are faulted remains of Santa Cruz (Bow, 1979). Platform Series lavas are both subaerial and submarine, thick (∼10 m) pahoehoe flows, and many are interbedded with beach and shallow water deposits (Bow, 1979). Field observations confirm that the Platform Series lies stratigraphically below the Shield Series wherever a contact between the units is exposed. Schwartz et al. (2022) identified Platform Series lavas in several locations (e.g., SC12-572B) along the North Puerto Ayora fault that extends NW from Santa Cruz’s largest town, along the south coast of the island. The 416 ± 36 ka (2σ) age of SC12-572B (Schwartz et al., 2022) is anomalously young for a Platform Series lava. Thus, the Platform Series may have been active as recently as ∼0.4 Ma (nearly concurrent with some Shield Series flows), and may not have shut down entirely at ∼1.1 Ma (White et al., 1993; Schwartz et al., 2022), but the lack of any other Platform Series samples from <1.1 Ma suggests that sample SC12-572B represents only a trivially small-volume tail of activity, and that the bulk of the Platform Series was emplaced between 1.6 and 1.1 Ma.

The Shield Series consists of predominantly pahoehoe flows erupted from the WNW-trending fissure system bisecting the island (Figure 1B; Bow, 1979). The fissures are manifested as a quasi-linear arrangement of late-stage cinder cones and pit craters in the highlands and tuff cones near the coast. Gravity studies by Cleary et al. (2020) confirm that they do not define the remnants of an in-filled caldera (Bow, 1979). Shield Series flows exhibit features typical of pahoehoe lava, with inflation lobes and tumuli.

Two Shield Series lavas have been dated using cosmogenic helium techniques, yielding ages between 0.111 ± 0.006 Ma and 0.585 Ma ± 0.013 (Kurz and Geist, 1999). On the basis of Ar-Ar ages and normal paleomagnetic polarity, Schwartz et al. (2022) estimates that the Shield Series became active ∼700 ka. Consequently, >1 million years of volcanic and geochemical history are preserved and exposed at Santa Cruz. Unfortunately, dense vegetation severely limits detailed field mapping across most of the island.

Along the northern, eastern, and southern coasts, a series of normal faults strike sub-parallel to the highland cinder cone lineation (Schwartz et al., 2022; Figure 1B). Schwartz et al. (2022) proposes that the faults define two distinct periods of tectonic activity. The older faults, limited to the northeast quadrant, only crosscut Platform Series lavas and formed since 1160 ± 35 ka, and likely between 780 and 500 ka (Bow, 1979; Schwartz et al., 2022), as a consequence of regional extension early in the island’s construction. The younger faults, along the southern flank, formed contemporaneously with the Shield Series, between 274 ± 18 and 38 ± 8 ka (Schwartz et al., 2022). Schwartz et al. (2022) attributes the younger faults to southward extension of the island caused by E-W-oriented magmatic intrusions that ceased ∼20 ka.

Samples were analyzed for major elements at Colgate University by X-Ray Fluorescence (XRF) spectroscopy following a method adapted from Norrish and Hutton (1969) on a Philips PW240 instrument. Replicate analyses of USGS Standard Reference Material (SRM) BHVO-2 yielded precision for all major element oxides <0.85% (1σ; Table 1). The precision of triplicate analyses of two samples (SC12-001 and SC12-063) yielded <1% relative standard deviation (RSD, 1σ) for all major elements, except P₂O₅ (1.2% RSD, 1σ). A subset of previously collected samples (Bow, 1979) was analyzed by XRF at Washington State University’s (WSU) GeoAnalytical Laboratory following procedures described in Johnson et al. (1999). Whole rock major element data were not adjusted for phenocryst content (e.g., crystal accumulation).

Trace element concentrations were determined via Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) using a Varian 820MS quadrupole instrument at Colgate University. A 200-ppb internal standard solution of ¹¹⁵In, ¹³³Cs, ¹⁸²W, and ²⁰⁵Tl was diluted to 1:20 using in-line mixing during analysis. Analyte masses were corrected to the nearest internal standard mass of those listed above to account for instrument drift, according to the method developed by Eggins et al. (1997). External standard curves were created using solutions of USGS SRMs BHVO-2, BIR-1, DNC-1, AGV-2, and W-2. The Pearson correlation coefficients of the standard curves were consistently >0.998. A solution of USGS SRM W-2 was analyzed every five to six samples as an unknown and yielded <4% RSD (1σ) precision for elements with atomic masses greater than 89. Only Sc and Rb registered precisions >8% (1σ) in replicate W-2 solutions run as unknowns; elements with masses less than 89 amu had precisions <8% and usually <5%. Corrections were made for isobaric interferences on Gd.

Strontium, neodymium, and lead radiogenic isotope analyses were performed at the University of Florida’s Department of Geological Sciences Table 3. Samples were leached aggressively to remove sea salt spray and weathering products according to procedures in Goss et al. (2010). All subsequent chromatographic elemental separations were performed in a class 100 clean laboratory. Samples were diluted with 2% HNO₃ to achieve ∼4–6 V total ion current for each solution concentration. To monitor instrument performance, NBS 987 Sr and JNDi Nd standards were analyzed every four samples. External precision for Sr (2σ) is 0.00002 (n = 42) and 0.00002 (n = 91) for Nd. Strontium results were corrected to a ⁸⁸Sr/⁸⁶Sr value of 0.1194 and neodymium ratios were normalized to a ¹⁴⁶Nd/¹⁴⁴Nd value of 0.7219 using an exponential law for mass bias fractionation. Strontium and neodymium isotope ratios presented here have been corrected to SRM NBS 987 Sr (⁸⁷Sr/⁸⁶Sr = 0.710248) and La Jolla (¹⁴³Nd/¹⁴⁴Nd = 0.511858). The Pb standard, NBS 981, achieved a long-term (June 2012–January 2013) average of ²⁰⁶Pb/²⁰⁴Pb = 16.938 ± 0.004 (2σ), ²⁰⁷Pb/²⁰⁴Pb = 15.488 ± 0.003 (2σ), and ²⁰⁸Pb/²⁰⁴Pb = 36.693 ± 0.009 (2σ). The Pb isotopic analyses presented here are corrected to NBS 981 values ²⁰⁶Pb/²⁰⁴Pb = 16.9405, ²⁰⁷Pb/²⁰⁴Pb = 15.4963, and ²⁰⁸Pb/²⁰⁴Pb = 36.7219 (Weis et al., 2006). A subset of Santa Cruz samples were analyzed at the Pacific Centre for Isotopic and Geochemical Research (PCIGR) at the University of British Columbia, according to methods described in Harpp and Weis (2020). All data have been corrected to the same reference values as samples analyzed at the University of Florida.

According to thin section observations, lavas in the Platform Series are aphyric to plagioclase-phyric with a subophitic groundmass composed of plagioclase, clinopyroxene, olivine, and opaque oxides. Plagioclase phenocrysts are up to 1 cm long, and constitute 1–3% of Platform Series samples. A plagioclase ultraphyric lava in the Cerro Colorado area contains 25% modal plagioclase, with crystals up to 10 × 5 mm, but its major element composition is not exceptional compared to the others from the same unit. Most lavas have sparse (<2%) olivine phenocrysts ∼1 mm in diameter; however, olivine and plagioclase are often present in glomerocrysts up to 1 cm in diameter.

Shield Series lavas include plagioclase- and olivine-phyric rocks. Plagioclase and Ti-rich clinopyroxene are ubiquitous groundmass phases. Olivine is also present in the intergranular to subophitic groundmass. Plagioclase phenocrysts are variable in shape and typically between <1 and 5 mm long. They constitute up to 25% modal abundance in some samples, although the majority have <10% phenocrysts. Olivine phenocrysts are also prevalent in Shield Series lavas, usually <25% modal abundance. Olivine phenocrysts are subhedral to euhedral, ∼1–2 mm in diameter, and some contain spinel inclusions. Hawaiites erupted at cinder and tuff cones preserve a pilotaxitic matrix rich in plagioclase microlites, oxidized olivine, and opaque oxides all <1 mm long.

Major element concentrations of Santa Cruz lavas define a wide compositional range that spans values observed across the archipelago (Figures 2–4; Table 1; e.g., White et al., 1993), despite a relatively small range of SiO₂ contents (45.1–49.6 wt.%); rocks are tholeiitic to weakly alkaline (Figure 2). Major element variations do not exhibit any systematic geographic pattern. Santa Cruz lavas have between 4.3 and 12.9 wt.% MgO, extending to both more primitive and evolved compositions than Fernandina and most of the other active volcanoes in the western Galápagos (Figures 3A–H; Allan and Simkin, 2000). At a given Mg# value, Santa Cruz has lower TiO₂ and CaO contents than Fernandina, but higher Al₂O₃, FeO_t, Na₂O, and P₂O₅. In general, K₂O values at a given Mg# are higher than those from Fernandina for most of the Shield series, but lower for the Platform Series.

The Platform Series rocks are more fractionated than Shield lavas, but cannot be produced from the more primitive Shield Series samples, owing to their different incompatible element ratios and the distinct trends they define in Harker diagrams. The lack of dominant trends in either the Platform or Shield Series lavas suggests that there is not a single parental melt composition responsible for each geologic unit; instead, there must be an array of primary melt compositions and depths of melt generation within both the Shield and Platform Series. Similarly, the range of FeO_t contents in primitive lavas may signal that the mantle source includes a pyroxenitic component, as has been proposed for other lavas from the central Galápagos (Gleeson et al., 2020).

In CaO-Mg# space (Figure 3E), Santa Cruz lavas define two broad trends. The positive trend that includes most of the Shield Series lavas likely reflects control primarily by clinopyroxene fractionation, whereas the shallower trend defined by Platform lavas and a handful of mostly low K₂O/TiO₂ Shield lavas has experienced less clinopyroxene fractionation. The high modal abundance of plagioclase and elevated Al₂O₃ contents of many Shield Series lavas indicates that there has been plagioclase accumulation. In contrast, the Platform Series’ lower Al₂O₃ contents suggest greater control by plagioclase fractionation.

The Platform Series defines a horizontal array in Mg#-K₂O/TiO₂ space, varying between ∼0.1 and 0.25 in K₂O/TiO₂. The low FeO and MgO values of some Platform Series suggest that a few may have fractionated Fe-Ti oxides, although to a limited extent given that these lavas maintain relatively elevated TiO₂ contents (Figures 3B,D). The Shield Series array is negatively sloped in Mg#-K₂O/TiO₂ space; consequently, we have divided the Shield Series into three compositional sub-groups (Figure 4). Shield Series lavas with K₂O/TiO₂ <0.15 are classified as the Low K₂O/TiO₂ suite. These samples have the highest MgO and CaO, and the lowest K₂O, TiO₂, Na₂O, and P₂O₅, of the Shield Series (Figures 3A,B,E–H). Samples with K₂O/TiO₂ >0.25 are the High K₂O/TiO₂ suite, and constitute the most evolved lavas within the Shield. Lavas with K₂O/TiO₂ values between 0.15 and 0.25 are designated as the Mid K₂O/TiO₂ suite of the Shield Series (Figure 4).

Broadly, trace element concentrations vary from depleted, near-MORB-like values to more enriched compositions typical of ocean island basalt (Table 2; White et al., 1993). The Platform Series exhibits less compositional variability in trace element contents than the Shield Series (Figures 5–7).

Basalts from Santa Cruz define a wider range of isotopic ratios than those erupted at most other Galápagos Islands (Figures 8A–D; Table 3, e.g., White et al., 1993). The ⁸⁷Sr/⁸⁶Sr ratios vary from 0.702633 to 0.702959, and ε_Nd values extend from 7.77 to 9.43. Lead isotopic ratio ranges are 18.531–18.947 for ²⁰⁶Pb/²⁰⁴Pb, 15.519–15.560 for ²⁰⁷Pb/²⁰⁴Pb, and 38.071–38.515 for ²⁰⁸Pb/²⁰⁴Pb. Platform Series lavas are generally more enriched than the Shield Series. Even though there is overlap between the two groups, taking into account analytical uncertainty, the averages of the Platform and Shield Series lavas in all isotopic ratios are significantly different from each other at >95% certainty. The enrichment in Platform over Shield Series lavas supports previous observations based on fewer samples (White et al., 1993).

The most depleted signatures of the Shield Series (e.g., ⁸⁷Sr/⁸⁶Sr < 0.7027; ε_Nd > 9) resemble material erupted at Genovesa Island in the Northern Galápagos Volcanic Province (NGVP; e.g., Harpp et al., 2003; Harpp et al., 2014c). The most enriched Santa Cruz lavas are comparable to the enriched material erupted at other eastern Galápagos volcanoes, including Santiago and San Cristóbal, which also display comparable ranges in isotopic ratios (e.g., Geist et al., 1986; Gibson et al., 2012). Santa Cruz lavas are not as enriched as those produced at volcanic centers closer to the leading edge of the plume, such as Fernandina, Volcán Sierra Negra, and Volcán Cerro Azul (e.g., White et al., 1993; Allan and Simkin, 2000; Naumann et al., 2002).