Some of Earth's largest magmatic provinces result from the interaction between mid-ocean ridges and near-ridge hotspots, which are hypothesized to overlie plumes of upwelling mantle. Geodynamic models predict that upwelling plumes are sheared by the motion of the overlying tectonic plates and can connect to a nearby mid-ocean ridge by shallow flow beneath thin, young lithosphere. Here we present seismic tomographic images of the upper 300 km of the mantle beneath the Galápagos Archipelago in the eastern Pacific Ocean. We observe a low-velocity anomaly, indicative of an upwelling plume, that is not deflected in the direction of plate motion. Instead, the anomaly tilts towards the mid-ocean ridge at depths well below the lithosphere. These characteristics of the plume-ridge connection beneath the Galápagos Archipelago are consistent with the presence of multiple stages of partial melting, melt extraction, and melt remixing within the plume and surrounding mantle. These processes affect the viscosity of the asthenosphere, alter the upwelling plume and influence the compositions of surface lavas. Our results imply that the coupling between the oceanic plate and plume upwelling beneath the Galápagos is weak. Multistage melting may similarly affect the geophysical and geochemical characteristics of other hotspots.

[1] We present a Rayleigh wave tomographic study of the upper mantle beneath the Galápagos Archipelago. We analyze waves in 12 separate frequency bands (8-50 mHz) sensitive to shear wave velocity (V S ) structure in the upper 150 km. Average phase velocities are up to 2 and 8% lower than for 0-to 4-My-old and 4-to 20-My-old Pacific seafloor, respectively. Laterally averaged V S is 0.05-0.2 km/s lower between 75-and 150-km depth than for normal Pacific mantle of comparable age, corresponding to an excess temperature of 30 to 150°C and $0.5% melt. A continuous low-velocity volume that tilts in a northerly direction as it shoals extends from the bottom of our model to the base of a high-velocity lid, which is located at depths varying from 40 to 70 km. We interpret this low-velocity volume as an upwelling thermal plume that flattens against the base of the high-velocity lid. The high-velocity lid is $30 km thicker than estimated lithospheric thickness beneath the southwestern archipelago, above the main region of plume upwelling. We attribute the thicker-than-normal high-velocity lid to residuum from hot spot melting. The thickness of the lid appears to control the final depth of melting and the variability of basalt composition in the archipelago. At depths less than 100-120 km, plume material spreads in directions both toward and against eastward plate motion, indicating that plume buoyancy forces dominate over plate drag forces and suggesting a relatively high plume buoyancy flux (B ! 2000 kg/s).

[1] To constrain the seismic velocity structure of the crust beneath the Galápagos Archipelago, we conducted a tomographic study using high-frequency Rayleigh waves obtained from cross correlations of ambient noise. We analyzed waves with periods between 5 and 8.5 s, sensitive to shear wave velocity (V S ) structure between about 3 and 10 km depth, after accounting for the effect of water depth. Crustal velocities are up to 25% lower than those of very young crust at the East Pacific Rise and are comparable to those of Hawaii. We attribute the lower than normal velocities to the combined effect of heating and the presence of melt in the crust above the Galápagos plume as well as the construction of a highly porous volcanic platform emplaced atop preexisting oceanic crust. On average, V S between 3 and 10 km depth beneath the western archipelago is up to 15% lower than beneath the eastern archipelago. We attribute the west-to-east velocity increase to a decrease in porosity of the volcanic platform and to cooling of the crust after its passage above the Galápagos plume. The results of this study, in combination with previous work, indicate that many of the unusual aspects of the Galápagos Archipelago are the result of variations in the thickness and internal structure of the chemical and thermal lithospheres. Our findings indicate that observed variations in the flexural response to loading observed in the Galápagos cannot be explained by the current thermal state of the lithosphere. Instead, the flexural response likely represents varying elastic strength at the time of loading. We also propose that the northwest and northeast trending alignments of volcanic centers found throughout the archipelago (the Darwinian lineations) may be associated with preexisting zones of weakness in the lithosphere formed during earlier episodes of ridge jumping and ridge propagation that were later reactivated by stresses generated by plume-lithosphere interactions.

No abstract