DI41A-2587
Fine Scale Structure of Low and Ultra-Low Velocity Patches in the Lowermost Mantle: Some Case Studies

Thursday, 17 December 2015
Poster Hall (Moscone South)
Kaiqing Yuan, Berkeley Seismological Lab, Berkeley, CA, United States, Barbara A Romanowicz, University of California Berkeley, Berkeley, CA, United States and Scott French, NERSC, Oakland, CA, United States
Abstract:
The lowermost part of the mantle, which is roughly halfway to the center of the earth, plays a key role as a thermal and chemical boundary layer between the solid, silicate mantle and fluid, iron outer core. Constraining the seismic velocity structure in this region provides important insights on mantle dynamics, and core-mantle interactions.

Recently, global shear wave velocity tomography has confirmed the presence of broad plume conduits extending vertically through the lower mantle in the vicinity of major hotspots (SEMUCB-WM1, French and Romanowicz, 2015). These conduits are rooted in D" in patches of strongly reduced shear velocity, at least some of which, such as Hawaii, appear to contain known ultra low velocity zones (e.g. Cottaar and Romanowicz, 2012). We seek to determine whether these patches generally contain ULVZs, and to contrast them with less extreme structures such as the PERM anomaly (Lekic et al., 2012). Because global tomography cannot resolve such fine scale structure, we apply forward modeling of higher frequency (10-20s) Sdiff waveforms in 3D complex structures using the Spectral Element Method. We focus on Iceland, Hawaii and the PERM anomaly, and Sdiff observations at USArray and/or dense broadband arrays in Europe. In all three cases, Sdiff waveforms are clearly distorted by these anomalies, with either a complex coda and/or evidence for amplitude focusing. As a start, we design simple cylindrical models of shear velocity reduction, and contrast the best fitting ones at each location considered in terms of diameter, height above the core-mantle boundary and strength of velocity reduction. We refine previously obtained models for Hawaii and the Perm Anomaly. For Iceland, the waveforms show a strong azimuthally dependent post-cursor, with maximum travel time delay of ~20s and focusing effects. The preliminary best fitting model shows a structure of 700km in diameter, ~15% reduction in shear wave velocity, extending ~40 km above the core-mantle boundary, in a location close to the Iceland hotspot which is in agreement with the low velocity patch in model SEMUCB_WM1.