DI005-0014
New large-scale ultralow-velocity zones discovered using highly anomalous SPdKS recordings
New large-scale ultralow-velocity zones discovered using highly anomalous SPdKS recordings
Wednesday, 9 December 2020
Poster
Abstract:
We present a new global map of ultralow-velocity zone (ULVZ) locations using a global collection of highly anomalous SPdKS recordings. We analyzed a global dataset of more than 58,000 radial component seismograms in the epicentral distance range from 106° to 115°, searching for highly anomalous SPdKS recordings that may be indicative of ULVZ presence. These data sample 56.9% of the CMB by surface area and provide the most comprehensive search for ULVZs to date. Because of the inherent source-/receiver-side path ambiguity between the complementary SPdKS/SKPdS seismic phases we conducted an inversion for the most likely ULVZ locations, using the principle of parsimony. The inversion is conducted using a genetic algorithm, and due to non-uniqueness of possible solutions, we repeated the inversion several thousand times in order to construct a ULVZ probability map. This analysis reveals on the order of 26 regions with high probability of ULVZ existence that explain the majority of highly anomalous SPdKS waveforms. These regions encompass roughly 10% of the surface area of the core-mantle boundary explored. The regions with the highest ULVZ probability include regions already suggested to contain ULVZs such as beneath central Africa, the coral Sea, Samoa, and northern Mexico. But, roughly 10 new ULVZs are also revealed such as beneath northern Africa, East Asia, South America, and British Columbia. Some of these newly discovered ULVZs are potentially as large in size as the most massive ULVZs discovered to date and may be of a similar size as those inferred beneath Samoa, Hawaii, and Iceland. In order to confirm the results of our inversion, we analyze ScP recordings interacting with a newly inferred ULVZ located to the east of the Philippine Islands, demonstrating that the presence of a large ULVZ in this region is also consistent with ScP data.