Transition-zone structures derived from USArray triplication data

Recent deployment of the USArray has provided an extraordinary dataset to study seismic structures beneath a large portion of the North American continent. In this dataset, earthquake data with distance larger than 30° has been extensively explored by various inversion techniques, such as imaging of teleseismic arrivals with tomography, shear-wave splitting mapping, receiver function analysis, etc. The results, however, have relatively low resolution on deep structures below 300 km outside the array (Chu and Helmberger, 2011). On the other hand, less attention has been paid to regional earthquake data between 10° and 30°, which has relatively high resolution. The reason is mainly due to the complexities of the upper-mantle triplications from the 410 and 660 discontinuities. In this study, we will exploit the differential travel times between the CD and EF branches of the upper-mantle triplication at distance of 23° to 30° to study the transition-zone velocity structures. Seismic waves of the CD and EF branches turn in the transition zone and in the lower mantle, respectively, and share similar ray paths in the upper mantle. Therefore the differential travel times between the CD and EF branches are sensitive to the velocity structures near the turning points. We apply the recently developed multi-pathing detector (Sun and Helmberger, 2011) to accurately measure the differential times. The analysis of regional data from earthquakes along the west coast, from central Alaska to central Mexico, and in the Gulf of Mexico will be presented. Preliminary results from an earthquake in the Gulf of Mexico sampling the northern gulf and Louisiana show that the differential times between CD and EF are 2.0-7.0 s larger than synthetics of SNA. We attribute this larger separation to a faster EF branch, which indicates that the Farallon slab may be shallow at about 700 km, as displayed in recent tomographic images (Simmons et al., 2010).