80% of earthquakes occur underwater, so ocean-bottom seismometers (OBS) are crucial for improving our understanding of earthquake source mechanics along unexplored offshore faults, and fillling key gaps in our images of the deeper solid Earth. Even for ocean islands and island arcs, land stations alone struggle to image underlying structures. Broadband OBSs have been through many design iterations, but many OBS deployments now yield high data recovery rates (>90%).
Even though my first OBS deployment experience left me feeling seasick, I have since continued to seismically explore the oceans, taking part in several OBS projects. In this talk, I will focus on my recent results from experiments across the Atlantic Ocean. Compared to the faster-spreading and subducting Pacific lithosphere, the less well-studied Atlantic offers a key endmember for refining our knowledge of global tectonics and associated hazards.
In the Lesser Antilles subduction zone, subducting Atlantic lithosphere is heterogeneously hydrated. Local earthquakes recorded by OBSs (VoiLA experiment), allowed me to image seismic attenuation to map fluid and melt pathways through the slab and mantle wedge, showing how slab fluids precondition melt generation and volcanism in arc settings. In the mid-Atlantic, long transform faults can host large M~7 earthquakes in ultra-wide (20-30 km thick) fault zones, allowing a uniquely macro-scale view of how damage zones control seismogenesis. In 2017, OBSs (PI-LAB experiment) recorded a nearby Mw 7.1 earthquake on the Romanche transform fault, triggering detailed teleseismic analysis that show back-propagating rupture fronts, which have since been seen during the 2023 M7.8 Türkiye earthquake. More recently, I analysed a seismic swarm and dyke intrusion in the Azores, which lies on a diffuse transtensional plate boundary. Here, a temporary OBS network (UPFLOW project) installed around the uniquely narrow island of São Jorge yields high-resolution seismicity locations that shed light on magma inflow and drainage along pre-existing faults.
Overall, OBS experiments yield fascinating results, but these results come from vast team efforts, particularly from ship crews and OBS technicians, that often go uncredited. We need to work harder to ensure the long-term sustainability of data from these expensive, often publicly-funded projects, with OBS-specific data preprocessing complications a partial barrier to this.