Chen Weiwen studied seismology at University of Science and Technology of China where he received a PhD in Geophysics in 2015. During his PhD career, he worked on the research of earthquake source parameters by seismologic and geodetic approaches. Since Oct. 2015, he is a research fellow at EOS , working on various seismology related projects, from global focal mechanisms to interior structure of the Earth.
Seismic tomography has been well-used for imaging the slab structure in the last couple decades and has greatly advanced our understanding of plate tectonics and geodynamics. However, regulation in the tomography inversions and relatively poor ray path coverage at depth hinder a higher resolution of velocity structure imaging, which is required to address one of the fundamental questions: why some of the slabs float above the lower mantle (660 km), while others can penetrate into it? Modelling of the high-frequency global waveform records provides complementary constraints to the details of the slab structure that can be used together with tomography models to shed some new lights on this issue.
In this talk, I will introduce a high-frequency waveform modelling procedure and its application to analyse the relative source time functions of the 2013 Mw8.3 deep Sea of Okhotsk earthquake, which was located inside of the North Kurile slab at the depth of ~620km. By analyzing sub-events on the high-quality USArray data in distant and azimuthal record-sections, we found that the second sub-event of the rupture was located at a depth 19.7km shallower than the first sub-event (and offset horizontally towards 37 °N by 6.2km). The spatial difference of these two subevents produced appreciable differences in their P-wave ray paths that sampled approximately along the North Kurile slab and well recorded by the EU array.
We adopted a 2D finite-difference algorithm to model the distant record-sections at the EU array with focus on the relative arrival time differences between the two sub-events. Our modelling results indicate that the North Kurile slab, which penetrates to ~1000km in depth with a dip angle of ~45°, has a thickness of 60km and up to 5% velocity perturbation, in contrast with the blurry tomography images in the region. Our results suggest the slab has been chopped off after penetrating through 660-discontinuity. The thinning of the slab in the lower mantle sparkles a new model to explain the tomography observation of Kurile slab system: the buoyant structure below 660-discontinuity filtered the lighter materials of the slab, and only material with higher density can pass through. With the extension of this trapped megalith and the detachment of the deeperhigh-densityy thin portion, it will eventually float again above the 660km discontinuity, just like its neighbour South Kurile slab, which is 10 million years older.