Abstract

Much of our understanding of tectonic processes on Earth come from studies focussed in the Pacific Ocean and along its margins. The impact of slower spreading rates in the Atlantic, and its impact on subduction and earthquake rupture processes has received less attention. My talk will focus on new results from two UK-led ocean-bottom seismometer experiments in the Atlantic during 2016-17: The VoiLA experiment in the Lesser Antilles subduction zone and the PI-LAB experiment on the mid-Atlantic ridge.

In subduction zones, the volume of water stored in the down-going plate and how it is released into the asthenosphere wedge is key for understanding the mass balance of Earth’s mantle. The Lesser Antilles subduction zone is a global end-member for slow subduction of slow-spread, hydrated oceanic lithosphere, which could cause variable hydration and melting of the mantle. Using local earthquakes, we have generated 3-D images of seismic attenuation, sensitive to mantle temperature and fluid content. We find the asthenosphere wedge does not lie beneath the arc as expected from thermal-mechanical modelling. Melt may pond at the base of the upper plate in the back-arc with lateral migration of melt through the crust towards the arc. The location of the hydrated core of the mantle wedge agrees well with geochemical estimates of water content from erupted magmas at the surface.

The simple geometry of ocean transform faults that offset spreading centres offer a rare opportunity to study how fault zone friction affects earthquake rupture. We recorded a magnitude 7 earthquake along the Romanche transform fault in the central Atlantic on ocean-bottom seismometers. Using these data along with teleseismic waveforms, we show that the rupture direction along the fault reversed mid-way through the earthquake. We theoretically expect back-propagating rupture fronts to occur in mature, low-velocity fault zones, but we have not observed before such clear evidence of a reversing rupture. This phenomenon is absent in rupture simulations and unaccounted for in hazard assessments.