Earthquake Processes along Oceanic Transform Faults and Ridge Segments: OBS Observations along the Chain Fracture Zone in the Mid-Atlantic from the PI-LAB Experiment.

Oceanic transform faults and ridge segments make up a large proportion of the world's plate boundaries. Our understanding of the physics of seismic slip along these plate boundaries remains poorly understood, and it is largely based on hypocenter locations and rupture models derived from data recorded at large epicentral distances. Based on teleseismic observations of larger earthquakes, the maximum depth of seismic rupture along the Mid-Atlantic transforms is believed to be thermally controlled, due to a proposed brittle-ductile transition at 600˚C. To present the first comprehensive high-resolution view of earthquake processes along Atlantic transform faults, we present a new catalogue of local seismicity and rupture source models around the Chain transform fault. Data comes from a dense (min. station spacing of 50 km) temporary network of 39 broadband ocean-bottom seismometers (OBS) covering a 12-month period as part of the PI-LAB project.

During the one-year deployment 2,600 potential events were automatically detected. Manual picking and iterative relocation was used to refine the catalogue, leaving 1,000 seismic events, with a minimum local magnitude of 2.0. The hypocenter locations were relocated using oct-tree cells in Non-Linear Location method (NonLinLoc) based on high-quality P- and S-wave arrival times. We find that the majority of earthquakes are located along the active transform faults in an area of 10 - 15 km width. A second set of events, significantly smaller in number, can be observed along the MAR and the fracture zones, which mark the inactive continuations of the transform faults. A third set, comprised of intra-plate locations, cannot be attributed to these tectonic features. While most of the earthquakes along the transform faults are situated shallower than 20 km, a significant number of deeper hypocenters can be found at depths of 30 - 40 km or deeper, particularly beneath the ridge-transform intersections; these depths are confirmed by regional full-waveform moment tensor inversions for the larger events (Mw > 4.5). These depths are well beyond the location of the 600˚C isotherm in those areas. We therefore discuss alternative mechanisms for such deep seismogenesis, including the possible role of hydrated mantle beneath the plate boundary.