Linking the Wrangellia flood basalts to the Galápagos hotspot

The Triassic volcanic rocks of Wrangellia erupted at an equatorial to tropical latitude that was within 3000 km of western North America. The mafic and ultramafic volcanic rocks are compositionally and isotopically similar to those of oceanic plateaux that were generated from a Pacific mantle plume-type source. The thermal conditions, estimated from the primitive rocks, indicate that it was a high temperature regime (TP > 1550 °C) consistent with elevated temperatures expected for a mantle plume. The only active hotspot currently located near the equator of the eastern Pacific Ocean that was active during the Mesozoic and produced ultramafic volcanic rocks is the Galápagos hotspot. The calculated mantle potential temperatures, trace elemental ratios, and Sr–Nd–Pb isotopes of the Wrangellia volcanic rocks are within the range of those from the Caribbean Plateau and Galápagos Islands, and collectively have similar internal variability as the Hawaii-Emperor island chain. The paleogeographic constraints, thermal estimates, and geochemistry suggests that it is possible that the Galápagos hotspot generated the volcanic rocks of Wrangellia and the Caribbean plateau or, more broadly, that the eastern Pacific (Panthalassa) Ocean was a unique region where anomalously high thermal conditions either periodically or continually existed from ~ 230 Ma to the present day.

Paleogeography, composition and thermal history of the Wrangellia flood basalt. The paleogeographic location of Wrangellia is constrained by various paleomagnetic studies (Fig. 2a). Yole and Irving 43 investigated the paleomagnetism of the Karmutsen Formation and concluded that the paleolatitude of Wrangellia is 18 ± 6° S or 18 ± 6° N. This would imply a paleolatitude of 12° S to 24° S or 12° N to 24° N. They also compared the results to those from Schwarz et al. 44 that determined paleolatitudes of 13 ± 15° and 17 ± 12° (which would be from 2° S to 39° N). Stone 45 based on paleomagnetic mean poles suggests that Wrangellia was likely very close to, and probably north of the equator. Panuska 46 also summarized that Wrangellia occupied a northern hemisphere position (10°-20° north latitude) during the late Paleozoic and early Mesozoic. Symons 47 , using paleomagnetic data from 46 sites (674 specimens) of the Crystalline Gneiss Complex on the west coast of Vancouver Island, indicates that the Wrangellia terrane was located at an 18 ± 6° S paleolatitude (i.e. 24° S to 12° S). Hillhouse and Gromme 48 measured 46 Triassic lava flows to yield a mean paleolatitude of 13° N or 13.9° S. They also summarized that extensive sampling of the Triassic rocks of Wrangellia in Alaska and British Columbia has consistently yielded paleolatitudes of 10°-17°. Hillhouse and Coe 49 summarized all the paleomagnetic data from Alaska (Wrangellia) and stated that "all parts of Wrangellia apparently originated within 18° of the Triassic equator". The prevailing paleogeographic reconstructions of the early Mesozoic (ca. 230 Ma), e.g. Golonka 50 , Scotese and Schettino 51 , and Cao et al. 52 , all place Wrangellia close to the equator.
The flood basalts of Wrangellia (Nikolai and Karmutsen lavas) are variably altered but most are tholeiitic with a minor amount of mildly alkalic flows 13,[15][16][17][18] . The Mg# of the basalts ranges from ~ 72 to ~ 27 indicating some rocks are compositionally primitive or near primitive (Mg# = 72-65) whereas others are differentiated (Mg# < 65). The basalts can be subdivided into high-Ti (≥ 1.4 wt%) and low-Ti (< 1.0 wt%) groups that reflect different sources 16 . Furthermore, high-Mg (MgO = 8 to 12 wt%; Mg# = 61-74) basalt and picrite (MgO = 13 to 20 wt%; Mg# = 70-78) are identified within the volcanic successions of the Karmutsen Formation. The high-Mg basalt and picrites are tholeiitic and have low-TiO 2 (< 1.0 wt%) concentrations, and testify to a high temperature regime 18,53 . The primitive mantle normalized La/Yb PM (0.4 to 12.9; avg = 2.1) and Sm/Yb PM (0.6 to 9.8; avg = 1.5) ratios of all rock types are variable but generally low (Fig. 3a). Moreover, their ΔNb values (ΔNb = 1.74 + log[N b/Y] − 1.92log[Zr/Y]), an indicator of source characteristics, are nearly all > 0 (ΔNb = − 0.11 to 0.71; avg. = 0.16) and consistent with a mantle plume source 56 (Fig. 3b) (Fig. 3). It is thought that the low-Ti tholeiitic basalt is derived by small degrees (< 5%) of melting of  Spatial correlation of Wrangellia, the Caribbean plateau, and the Galápagos hotspot. The mafic and ultramafic rocks of Wrangellia are considered to be derived from a plume-type source that is similar to those which generated oceanic plateaux of the Pacific Ocean basin. The principal issue however, is the loca-  www.nature.com/scientificreports/ tion of the hotspot that generated the volcanic rocks as the pre-and post-volcanic carbonate rocks of Wrangellia contain fossils that are typical of tropical-equatorial latitudes to mid-latitudes of the eastern Pacific (Panthalassa) Ocean [58][59][60][61] . Paleomagnetic data from different sections of Wrangellia consistently yield equatorial to near equatorial latitudes (± 18° of the equator) at the time of their eruption 47,48,[50][51][52] . Furthermore, Belasky et al. 58 suggests that Wrangellia was within 2000-3000 km from coastal North America during the Early Permian based on fossils and the location of the East Pacific (Panthalassa) barrier (EPB) but, it is likely that at the time of eruption during the Carnian-Norian that Wrangellia was even closer to coastal North America than during the Permian. Torsvik et al. 62,63 and Burke and Torsvik 64 argue that the majority (~ 80%) of oceanic hotspots and continental flood basalt provinces emplaced since the Carboniferous are spatially correlated to long term stability of the 1% slow-velocity contour in the lowermost layer of the mean shear-wave tomographic model (SMEAN). The 1% contour defines a plume generation zone and is referred to as a large low shear velocity province (LLSVP). The current LLSVPs are primarily located beneath the African plate and the Pacific plate. It is at the boundary regions of an LLSVP that thermally anomalous upwelling of deep-seated mantle is thought to occur and manifests at the surface as oceanic islands/plateaux and continental flood basal provinces 62,63 . If this is the case, then the paleogeographic location of the Wrangellia flood basalts can be constrained by superimposing the current Pacific LLSVP on a Carnian plate reconstruction map of Pangea as the African LLSVP was too far to the east at the time. The intersection of the paleomagnetic-derived latitudinal range of Wrangellia with the Pacific LLSVP and the EPB is outlined on Fig. 2a. The distances obtained from the intersection point of the Pacific LLSVP range from ~ 10,000 km at the farthest point from western North America to ~ 4200 km at the closest point. The farthest intersection point overlaps with the modern location of Hawaii but the closest intersection point is still outside the estimated distance of Wrangellia proposed by Belasky et al. 58 . However, it is likely that the LLSVPs are not fixed and can wander [65][66][67][68] . Thus, the mostly likely location of the hotspot that generated the Wrangellia volcanic rocks is within the latitudinal variation but between the current easternmost point of the ~ 1% contour of the current Pacific LLSVP and the region advocated by Belasky et al. 58 (Fig. 2a). www.nature.com/scientificreports/ The only known and active hotspot that corresponds to the possible paleogeographic area of Wrangellia is the Galápagos hotspot (Fig. 2b) 69 . The Galápagos hotspot is located at the equator and just south of the active spreading centre separating the Cocos plate and the Nazca plate (Fig. 4). It is responsible for the present day Galápagos Islands and has been active for at least 20 million years as submerged volcanic edifices along the Cocos and Carnegie Ridges can be traced back to their point of origin 70 . The Galápagos hotspot is also linked to the generation of mafic (alkaline and tholeiitic) and ultramafic (komatiites) volcanic rocks of the Caribbean Plateau ( Fig. 4) at ~ 90 Ma and ~ 70 Ma but may stretch back to 140-110 Ma 72-80 . Some kinematic plate reconstructions suggest the Caribbean Plateau developed 1000-3000 km east of the Galápagos hotspot whereas others indicate there is a spatial-temporal correlation [81][82][83][84][85] . The correlation between the paleogeographic eruption location of the Wrangellia flood basalt and the current Galápagos hotspot is intriguing and offers a possible explanation for the eruption of the Karmusten picrites and some primary basaltic lava as they require mantle potential temperatures > 1550 °C which is indicated for some Caribbean plateau rocks 18,53,57,78,86 . Moreover, a long-lived Galápagos hotspot model is supported by the Early Cretaceous (~ 140 Ma) basalt along the Nicoya Peninsula of Costa Rica and consistent with the development of the Caribbean Plateau by the accumulation of seamounts and oceanic plateaux at a subduction zone over a period of time rather than derivation by the initial plume head phase of the hotspot 75,78 .

SA
The eruption of ultramafic (picrite and komatiite) volcanic rocks during the Phanerozoic is relatively rare (e.g., North Atlantic Igneous Province, the Caribbean plateau, and Emeishan large igneous province) and they are all considered to be attributed to a mantle plume or hotspot 57,78,[87][88][89][90][91] . A comparison of the calculated mantle potential temperatures of Galápagos, Caribbean Plateau, and Wrangellia volcanic rocks using PRIMELT3 shows significant overlap but only the Caribbean Plateau and Wrangellia rocks extend to anonymously high estimates (Fig. 3c). Herzberg and Gazel 57 and Trela et al. 92 interpret the thermal decline from 90 to 70 Ma to the recent eruptions at the Galápagos Islands as evidence of a cooling trend in the hotspot related to elevated pyroxenite melt production.
Further support of the hotspot-association between the Wrangellia, Caribbean, and Galápagos volcanic rocks is their isotopic similarity. As previously noted by Greene et al. 16,18 , the total range of 87 Sr/ 86 Sr i , ε Nd ( t ), 208 Pb/ 204 Pb i , 207 Pb/ 204 Pb i , and 206 Pb/ 204 Pb i values of Wrangellia volcanic rocks overlap with those of the Caribbean Plateau and the Galápagos Islands ( Fig. 3d-h). Although the isotopic similarity cannot confirm ancestry from a specific mantle source or hotspot, it is still noteworthy that the isotopic compositions have a similar magnitude of internal variability as the rocks of the Hawaii-Emperor island chain 93 . Nevertheless, the elevated mantle potential temperatures, similar paleogeographic eruptive locations, and the rarity of Phanerozoic ultramafic lavas are compatible with a single, albeit isotopically heterogeneous, source hypothesis 16-18,53,57,72,90,92,94,95 . The eastern Pacific hotspot region and possible Mesozoic hotspot track. The longevity of magmatism at an oceanic hotspot is unknown but the Hawaiian hotspot has likely been active for 100-150 million years 96 . Furthermore, the Louisville and Arago (Rurutu) hotspots may have been active for ~ 120 million years as well 97,98 . Mantle plume tracks within continental crust suggested for the Mongolia plume of Central Asia (~ 120 m.y.) and the Great Meteor hotspot track (~ 200 m.y.) of North America both exceed 100 million years [99][100][101] (Fig. 5). The timeframe between the eruption of the youngest volcanic rocks of Wrangellia (~ 225 Ma) to the oldest rock (~ 140 Ma) considered to be related to the Caribbean plateau is ~ 85 million years 12,75 and within the known lifespans of active hotspots but is also within the range of the Great Meteor hotspot track if extended to include the modern Galápagos Islands, and the Cocos and Carnegie ridges.
Such a large time gap and the orientations of the rocks of Wrangellia and the Caribbean plateau is inconsistent with the 'continuous' creation of oceanic islands and seamounts that is typified by the Hawaiian hotspot but also observed through the magmatic spatial-temporal progression of the Great Meteor hotspot track. There are three possibilities that can explain the apparent lack of a magmatic track that would 'connect' the Wrangellia and the Caribbean plateau rocks: (1) there is no track and the Wrangellia and Caribbean rocks are 'unconnected' and derived from temporally distinct hotspots that developed within the same geographical region, (2) the 'missing' island track was subducted, or (3) there is a track but, is has yet to be identified. The different orientations of the Wrangellia (north-south) rocks and Caribbean (west-east) rocks could be related to ridge jump, hotspot drifting, or both 62,102-106 .
The hypothesis that the Wrangellia and Caribbean plateau rocks, and by association the Galápagos Islands, are unrelated to the same hotspot and that no track was created is reasonable and perhaps the most likely scenario. However, there are two implications for the 'unconnected' hypothesis that would be unusual for oceanic hotspots. Firstly, the volcanic rocks of Wrangellia erupted over a short period from ~ 230 to ~ 225 Ma 12 . There is nothing unusual with such a short eruptive duration per se as it similar to some continental large igneous provinces but, nearly all oceanic hotspots have island chains that indicate long-lived magmatism, plume migration, and plate motion 67,69,[107][108][109][110] . Therefore, it is unlikely that the hotspot responsible for the Wrangellia magmatism was shortlived and did not have a track. Secondly, regardless of the duration of magmatism, the high mantle potential temperature estimates indicate that the eastern Pacific/Panthalassa Ocean has been a region of anomalously hot mantle upwelling periodically for ~ 230 million years as the region also witnessed the eruption of picritic and komatiitic lavas of the Caribbean plateau. Consequently, it would appear that the eastern Pacific/Panthalassa Ocean was unique in this regard.
The subduction of the hypothetical 'Wrangellia hotspot' island track is possible as is it known that seamounts and oceanic islands enter the subduction zones of the Costa Rica margin, Aleutian margin, and Izu-Bonin margin 111,112 . In this case, the 'missing' seamounts and islands related to Wrangellia would be emplaced on an oceanic plate that was destroyed during eastward subduction beneath North America. Although the complete subduction of the island track is a possibility, many seamounts and oceanic islands are accreted to continental   www.nature.com/scientificreports/ margins and commonly identified in collisional belts 113,114 . Thus, the circumstances that led to the accretion of Wrangellia to North America and not the associated island chain requires an explanation. One such explanation could be that Wrangellia was built upon an older, relatively buoyant substrate (e.g., arc basement), whereas the island track was built directly upon oceanic crust 13 . Regardless, if the island chain subducted then the verification of its existence is problematic. The third possibility is that there are uncorrelated units of the 'Wrangellia hotspot' track that accreted to western North America. From northern Washington to southern California there are a number of Jurassic ophiolitic units that have reported ages ranging from ~ 190 to ~ 160 Ma and include the Ingalls Ophiolite (Washington), Oregon Coast Range Ophiolite (Oregon), Josephine Ophiolite (Oregon-California), and the Coast Range Ophiolites of California 54,115,116 (Fig. 6). The age of the northern ophiolites appears to decrease southward as the oldest reported age from the Ingalls Ophiolite  (Fig. 7).
The tectonomagmatic origins of the Jurassic ophiolites are a topic of considerable debate as there are three principal models proposed to explain their origin 54,115,116,120,121 . Ingersoll 120 summarizes the tectonomagmatic models of the Coast Range Ophiolites and offers arguments in favour and against each one. The models are: (1) "formation by intra-arc and back-arc spreading related to an east-facing intraoceanic arc" that collided with a westward oriented continental margin arc during the Kimmeridgian to Tithonian; (2) "formation by open-ocean seafloor spreading" and their subsequent "incorporation into the continent margin during trench initiation outboard of an existing continental-margin trench", and (3) "formation by forearc oblique rifting along the continental margin, followed by partial closure". Although the models are conflicting, all consider the ophiolites to be principally derived from oceanic lithosphere that developed by melting and emplacement at spreading centres and not hotspot related 115 .
The compositions of the basaltic rocks of the ophiolites are mostly similar to mid-ocean ridge basalt and island-arc tholeiites but there are within-plate compositions reported from the Ingalls and Coast Range ophiolites 54

Conclusions
The Triassic volcanic rocks of Wrangellia are considered to be derived from a Pacific-type mantle plume source. The exact location of the hotspot is uncertain but fossil and paleomagnetic data indicate that it was located at equatorial to topical latitudes of the eastern Pacific (Panthalassa) ocean. The paleogeographic location of the Wrangellia hotspot is within the region of the current Galápagos hotspot. A comparison of the mantle potential temperature estimates, trace element geochemistry, and Sr-Nd-Pb isotopes between the volcanic rocks of Wrangellia, Caribbean plateau, and the Galápagos Islands shows significant overlap and the geochemical variability is similar to other oceanic island chains (e.g., Hawaii-Emperor island chain). Our model necessitates that the potential Wrangellia-Caribbean-Galápagos hotspot was active for ~ 230 million years which is within the range of activity for the Great Meteor hotspot. The apparent absence of a confirmed hot spot track argues against a direct connection between the Late Triassic Wrangellia volcanic rocks and the Early Cretaceous initial flows of the Caribbean plateau but, it is possible that Early to Middle Jurassic oceanic islands/seamounts were either subducted or accreted to North America (e.g., Iron Mountain unit of the Ingalls Ophiolite). The evidence of a link between the volcanic rocks of Wrangellia, Caribbean plateau, and Galápagos Islands to a common, longlived hotspot is compelling and cannot be easily dismissed. The Wrangellia-Caribbean-Galápagos connection is possible and they are either related to a single, long-lived equatorial hotspot or, more broadly, that the equatorial region of the eastern Panthalassa/Pacific Ocean, near the Americas, has been a region of anomalously hot mantle upwelling for ~ 230 million years. If our model is correct, then the Galápagos hotspot is the longest continually active hotspot of the Phanerozoic. The primary issues that must be resolved are tighter paleomagnetic constraints on Wrangellia volcanic rocks and the discovery of more Early to Middle Jurassic rocks that are OIB-like within the North American Cordillera south of British Columbia.

Methods
Geochemical data of the Wrangellia, Caribbean plateau, and Galápagos Islands volcanic rocks was compiled using GEOROC (http:// georoc. mpch-mainz. gwdg. de/ georoc/) and can be found as supplementary table S1. The primary melt compositions and mantle potential temperature estimates were calculated using PRIMELT3 120 .
The major elemental data of each sample was entered into PRIMELT3 and calculated using an  www.nature.com/scientificreports/