The majority of Earth's seismicity is focused at subduction zones, either in the shallow thrust zone or as intermediate-depth (~50 to 300 km depth) or deep (to ~600 km depth) slab seismicity. Intermediate-depth seismicity is commonly thought to be caused by dehydration embrittlement, in which the dehydration of hydrous minerals in the subducted crust or mantle increases the pore fluid pressure in low-permeability rocks, allowing rocks to behave brittlely; although recent studies have called this hypothesis into question due to a lack up supporting laboratory observations. Double seismic zones are obvious in some subducting slabs (e.g., Honshu), but their presence in other slabs is less clear due to the relatively large errors in hypocentral locations relative to the spacing between the two layers of seismicity, which varies among different slabs. As the lower layer of seismicity is thought to be caused by the dehydration of serpentine in the slab mantle, having a limited view of the global prevalence of double seismic zones limits the constraints on slab mantle serpentinization and dehydration. In this project, we will relocate teleseismic earthquakes within 54 small arc sections and compare these seismicity distributions with pre-existing 2D thermal and mineralogical models, with a focus on the locations of greatest dehydration of slab minerals. The double-difference relocation technique teletomoDD will be used with nested regional-global 3D velocity models, which will significantly reduce scatter and improve the absolute locations of earthquake hypocenters in comparison to existing global catalogs. By better understanding intermediate-depth seismicity and its relation to thermal and petrological models, we can gain insight to the seismogenic processes in slabs and contribute to the understanding of six key questions pertaining to subduction zones: 1) What phenomenon is responsible for the majority of intermediate-depth seismicity? 2) Where are double seismic zones observable? 3) In double seismic zones, what is responsible for the lower layer of seismicity? 4) What controls the maximum depth of slab seismicity? 5) Is the distribution of slab seismicity correlated with changes in volcanism? 6) Is there a relationship between slab seismicity and changes in deep plate coupling?
The majority of Earth's seismicity is focused at subduction zones, where cold, dense oceanic crust and mantle descend beneath a more buoyant plate, with the ability of producing large earthquakes and volcanism. Seismicity occurs either in the shallow thrust zone, or as intermediate-depth (~50 to 300 km depth) or deep (to ~600 km depth) slab seismicity. The mechanisms that cause shallow earthquakes are relatively well understood, whereas the cause of intermediate-depth seismicity is less certain. This is due in part to the expected ductile behavior of Earth materials at high pressures and temperatures, conditions that are difficult to reproduce in the laboratory. However, by combining numerical models of temperature and mineralogical reactions within subducting crust and mantle to observations of seismicity, we can better understand what slab conditions are most closely associated with observed seismicity. In this project, we will relocate intermediate-depth earthquakes using techniques that more accurately reproduce the paths of seismic wave propagation and reduce errors in earthquake locations from existing global seismicity catalogs. We will then comprehensively compare the distributions of slab seismicity with existing temperature and mineralogical models to identify which particular reactions contribute most to the creation of potentially large intermediate-depth earthquakes.
Although most of the world's earthquakes occur relatively close to the surface (mainly shallower than 50 km), there are also earthquakes that are observed to occur down to nearly 700 km depth. The physical mechanism(s) responsible for these deep earthquakes, which occur within the lithosphere that is subducted into the mantle at trenches, is a subject of scientific debate. The prevailing hypothesis is that certain minerals in the subducting lithosphere undergo chemical transformations that release water, and that process can lead to instability in the form of earthquakes. In turn, the released water may play a role in generating magmatic activity at subduction zones. Our work applies state-of-the-art techniques to produce the most accurate estimation possible of the location of the earthquakes occurring in the subducting lithosphere. The configuration of slab seismicity, often in the form of what is termed double seismic zones (two nearly parallel layers of earthquakes), can provide constraints on thermal and chemical models of mineral dehydration. Our approach is to use the difference in travel times between nearby earthquakes and each seismic station that observes them to tightly constrain the relative locations of the earthquakes. This is most effective when we use computer alignment of seismic wave arrivals (cross-correlation) to determine the relative times. We find that this approach can lead to at least a 25% reduction in the scatter of the earthquake locations that are due to inaccurate arrival time estimates. Our results suggest that some regions where double seismic zones have been reported may be artifacts of poor location quality. This project has supported the research and professional development of two female Ph.D.-level researchers at UW-Madison. It has also fostered a collaboration between UW-Madison scientists and researchers at the University of Memphis and Southern Methodist University (including two female scientists), and has helped continue collaborations between UW-Madison and the University of Science and Technology of China.
